SentimentTestingUSDLoop.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "SentimentTestingUSDLoop.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/hugegene/LSTM-Prediction-of-Stock-Price-Movement/blob/master/SentimentTestingUSDLoop.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "dIF_jv42m9pH"
      },
      "cell_type": "markdown",
      "source": [
        "# Rolling LSTM testing on USD price movements"
      ]
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "hOXobOBeKwDe"
      },
      "cell_type": "markdown",
      "source": [
        "## 1 Objective and Overview\n",
        "\n",
        "**Objective, Overview**  \n",
        "A supervised learning was done to make trading decisions (Sell or Buy) for USD using USD's past price and MarketRisk TRMI. The supervised model\n",
        "used was LSTM.\n",
        "\n",
        "**Thoughts on the Results**  \n",
        "The LSTM model depends on a few hyerparameters that could affect the results:\n",
        "1. Train-Validate-Test Split\n",
        "2. Labels\n",
        "3. Features Selection\n",
        "4. Number of LSTM layers and nodes\n",
        "\n",
        "**Train-Validate-Test Split**  \n",
        "The LSTM seeks to remember consistent patterns from the training set. I think the splitting of train set is important in building a good model. A long\n",
        "train period may not produce better results as according to this experiment. If the dataset consist of random patterns, the LSTM model will not be able\n",
        "to replicate the training result on the validate or test set as the valildate or test set differs from the train set.\n",
        "\n",
        "**Target Labels**  \n",
        "The target labels used could affect the training. If the target labels were less random to its x variables, it could help the LSTM pick up the patterns\n",
        "easier. Engineering could be done to the target labels to help the training.\n",
        "\n",
        "**Features Selection**  \n",
        "Features that could predict the price movements would make the results better.\n",
        "\n",
        "**Number of LSTM layers and nodes**  \n",
        "After numerous testing, a 2 layer LSTM, each with 20 nodes would allow the model to converge if the model is able to converge.\n",
        "\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "mPEtY17IU0Qs"
      },
      "cell_type": "markdown",
      "source": [
        "## 1.1 Splitting Train-Validate-Test Set\n",
        "\n",
        "2007-01-01 data to 2018-12-31 USD price and TRMI MarketRisk data was used for this experiment.\n",
        "\n",
        "The data is split to train-validate-test set as below:\n",
        "\n",
        "![alt text](10 year train test split.JPG)\n"
      ]
    },
    {
      "metadata": {
        "id": "OjlMoZGeFd32",
        "colab_type": "code",
        "outputId": "772df02b-ab8c-4bec-f588-9f0e9a8383e7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 429
        }
      },
      "cell_type": "code",
      "source": [
        "# Plotting USD prices\n",
        "ax = tradeTable[\"2017-01-01\":\"2018-12-31\"].plot(y=['Close'], figsize=(12,6), grid=True)"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/pandas/plotting/_core.py:1716: UserWarning: Pandas doesn't allow columns to be created via a new attribute name - see https://pandas.pydata.org/pandas-docs/stable/indexing.html#attribute-access\n",
            "  series.name = label\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsYAAAFkCAYAAAAjVP3NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7fc285629a90>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "PCC2x8HuU4Y3"
      },
      "cell_type": "markdown",
      "source": [
        "## 1.2 Feature Engineering"
      ]
    },
    {
      "metadata": {
        "id": "tYnpeunFweP9",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "![alt text](https://github.com/hugegene/LSTM-Prediction-of-Stock-Price-Movement/blob/master/USDprice.png?raw=1)"
      ]
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "NHFZ6CcMVDaX"
      },
      "cell_type": "markdown",
      "source": [
        "USD price was overlayed with various TRMIs moving averages and the 30-days MarketRisk moving average seems to co-related well with the price visually. As such, the following was feature-enginnered as the input to the LSTM model for all the experiments:\n",
        "\n",
        "1. MarketRisk\n",
        "2. 7-day MarketRisk moving-average\n",
        "3. 30-day MarketRisk moving-average\n",
        "4. 90-day MarketRisk moving-average\n",
        "5. Returns from previous day Close (r1)\n",
        "6. Returns from previous 7th day Close (r2)\n",
        "7. Returns from previous 30th day Close (r3)\n",
        "8. Returns from previous 90th day Close (r4)\n",
        "\n",
        "## 1.3 Evaluation Metrics\n",
        "\n",
        "The evaluation metrics are:\n",
        "1. Matthews correlation coefficient (MCC)\n",
        "2. Accuracy\n",
        "3. Profit and Loss\n",
        "\n",
        "\n",
        "## 1.4 Labels\n",
        "\n",
        "Each timestep is labelled Sell(0) or Buy(1) depending on whether the next closing price is up or down.\n",
        "\n",
        "\n",
        "## 1.5 LSTM Set-Up\n",
        "\n",
        "The LSTM is set-up to predict the labels based on the sequential data of the past 5 time-steps.\n",
        "\n",
        "\n",
        "## 1.6 Trading\n",
        "\n",
        "Each LSTM prediction of Buy will result in a buying and holding of USD for 1 day  and each LSTM prediction of Sell will result in a short-sell and holding of USD for 1 day."
      ]
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "2c-AjZEwWonj"
      },
      "cell_type": "markdown",
      "source": [
        "## 2 Results "
      ]
    },
    {
      "metadata": {
        "id": "0qGLzi4Je4IT",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "EFF_TEST_START\n",
        "EFF_TEST_END"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "9X8agEoisJKp",
        "colab_type": "code",
        "outputId": "741f0338-57db-4301-dda6-b4dd887554e6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 239
        }
      },
      "cell_type": "code",
      "source": [
        "# calculating cumulative predicted and labelled trade returns over each time step\n",
        "PNLTable =  tradeTable.loc[EFF_TEST_START:EFF_TEST_END,:].copy()\n",
        "PNLTable[\"labelledPNL\"] =  (1+PNLTable['labelledreturns']).cumprod()\n",
        "PNLTable[\"PNL\"] = (1+PNLTable['predictedreturns']).cumprod()\n",
        "PNLTable.head()\n",
        "\n",
        "#plotting Confusion Matrix\n",
        "f = plt.figure(figsize=(15,7))\n",
        "ax = f.add_subplot(131)\n",
        "plotConfusion(EFF_TEST_START, EFF_TEST_END, \"\")\n",
        "\n",
        "f = plt.figure(figsize=(36,6))\n",
        "ax = f.add_subplot(131)\n",
        "\n",
        "# PNL Plot for the Test Period\n",
        "PNLTable[\"PNL\"].plot()"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-5-dbdf70b7dc2b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mPNLTable\u001b[0m \u001b[0;34m=\u001b[0m  \u001b[0mtradeTable\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mEFF_TEST_START\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mEFF_TEST_END\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mPNLTable\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"labelledPNL\"\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m  \u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mPNLTable\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'labelledreturns'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcumprod\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mPNLTable\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"PNL\"\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mPNLTable\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'predictedreturns'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcumprod\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mPNLTable\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhead\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'tradeTable' is not defined"
          ]
        }
      ]
    },
    {
      "metadata": {
        "id": "PI1G2mO9bdAE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 2.1 Evaluation\n",
        "\n",
        "The model overfit the validation set right from the start, making any improvement on the train set worse on the validation set. The most accurate model validated by the validation set predicts all Buy for the test period (as shown in the confusion matrix), making it not a good model."
      ]
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "p1nongBRVFzR"
      },
      "cell_type": "markdown",
      "source": [
        "##  3 Codes"
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "7qHBVmabJWCD",
        "outputId": "23d767ad-ae88-4f3a-889c-4b0c9e59d656",
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 108
        }
      },
      "cell_type": "code",
      "source": [
        "from google.colab import files\n",
        "uploaded = files.upload()"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-1f532a61-e35c-4e6e-9592-ee30fe478dc4\" name=\"files[]\" multiple disabled />\n",
              "     <output id=\"result-1f532a61-e35c-4e6e-9592-ee30fe478dc4\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Saving USDprices.csv to USDprices.csv\n",
            "Saving USDSentiment.csv to USDSentiment.csv\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "HGRXzvUR0X11",
        "outputId": "98471f3b-1b2a-4799-84d8-a7ed46625bfd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "!ls"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "sample_data  USDprices.csv  USDSentiment.csv\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "MxvJZZarlEe6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "d4d8aa4d-fe72-4b7c-a29e-13c3ee1b5830"
      },
      "cell_type": "code",
      "source": [
        "from math import sqrt\n",
        "from numpy import concatenate\n",
        "from matplotlib import pyplot as plt\n",
        "from pandas import read_csv\n",
        "from pandas import DataFrame\n",
        "from pandas import concat\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Activation, LSTM\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "from keras.callbacks import EarlyStopping\n",
        "from keras.models import model_from_json\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.metrics import accuracy_score\n",
        "from datetime import datetime\n",
        "import os\n",
        "import itertools\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "from sklearn.metrics import f1_score\n",
        "from sklearn.metrics import precision_score\n",
        "from sklearn.metrics import recall_score\n",
        "from keras.layers import Input, LSTM, RepeatVector, Lambda, Dense, Flatten, Permute, merge, multiply\n",
        "from keras.models import Model\n",
        "from keras import backend as K\n",
        "from sklearn.metrics import matthews_corrcoef\n",
        "from sklearn.metrics import roc_auc_score as areauc\n",
        "import tensorflow as tf\n",
        "from datetime import timedelta, date\n",
        "from dateutil.relativedelta import relativedelta"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "H0U6XhwlpraX"
      },
      "cell_type": "markdown",
      "source": [
        "## 3.1 Data Preparation"
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "CNggyVu2lcl4",
        "outputId": "9f5273ee-397e-45f7-8441-b8a3e1bc6479",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        }
      },
      "cell_type": "code",
      "source": [
        "##Read in Price data, use date as index\n",
        "Prices = pd.read_csv(\"USDprices.csv\")\n",
        "Prices.head()\n",
        "Prices['Date'].dtype\n",
        "Prices['Date'] = pd.to_datetime(Prices['Date'], format='%Y-%m-%d')\n",
        "Prices.head()\n",
        "Prices.index = Prices['Date']\n",
        "# print(Prices.head())\n",
        "\n",
        "#Reading in Sentiment Data and selecting News-only-MarketRisk Sentiment\n",
        "Sent = pd.read_csv(\"USDSentiment.csv\")\n",
        "Sent = Sent[Sent.dataType=='News'][['Date', 'marketRisk']].fillna(method = \"ffill\")\n",
        "# Sent.head()\n",
        "Sent['Date'] = pd.to_datetime(Sent['Date'], format='%Y-%m-%d')\n",
        "Sent.index = Sent['Date']\n",
        "# print(Sent.head())\n",
        "\n",
        "#Combining Price table and Sentiment Table\n",
        "Ana = Prices.merge(Sent, left_index= True, right_index=True, how='inner')\n",
        "Ana.head()\n",
        "Ana.columns\n",
        "Ana = Ana.drop(columns=['Unnamed: 0', 'Volume', 'Asset', 'Date_x', 'Date_y', 'Open', 'High', 'Low', 'UnadjClose'])\n",
        "Ana.head()\n",
        "\n",
        "# calculating r, %tage change of closing price over previous time step\n",
        "Ana['returns'] = (Ana['Close']- Ana['Close'].shift(1))/Ana['Close'].shift(1)\n",
        "Ana.head()\n",
        "\n",
        "# calculating target price\n",
        "Ana['Target'] = Ana[\"Close\"].diff(1).shift(-1)\n",
        "Ana.head()\n",
        "\n",
        "# calculating %change\n",
        "Ana['Change'] = Ana[\"returns\"].shift(-1)\n",
        "Ana.head()\n",
        "\n",
        "\n",
        "## Label each time step False(no-buy) or True(buy) based on whether the price will rise at the next closing price\n",
        "timestep = 1\n",
        "labels = []\n",
        "for i in range(0, Ana.shape[0]):\n",
        "  if(i+timestep< Ana.shape[0]):\n",
        "    aheadGain = [Ana[\"returns\"][i+j] for j in range(1,timestep+1)]\n",
        "    labels+= [np.sum(aheadGain)> 0]\n",
        "\n",
        "print(len(labels))\n",
        "\n",
        "# Price Table with Labels as Signal\n",
        "Ana = Ana.iloc[:-timestep,:].copy()\n",
        "Ana.head()\n",
        "Ana['Signal'] = labels\n",
        "Ana.head()\n",
        "\n",
        "#Feature Engineering\n",
        "Ana[\"Close30\"] =  Ana[\"Close\"].rolling(30).mean()\n",
        "Ana[\"Close100\"] =  Ana[\"Close\"].rolling(100).mean()\n",
        "Ana[\"r1\"] =  Ana[\"Close\"].diff(1)\n",
        "Ana[\"r2\"] =  Ana[\"Close\"].diff(7)\n",
        "Ana[\"r3\"] =  Ana[\"Close\"].diff(30)\n",
        "Ana[\"r4\"] =  Ana[\"Close\"].diff(90)\n",
        "Ana[\"marketrisk_avg30\"] = Ana[\"marketRisk\"].rolling(30).mean()\n",
        "Ana[\"marketrisk_avg90\"] = Ana[\"marketRisk\"].rolling(90).mean()\n",
        "Ana[\"marketrisk_avg7\"] = Ana[\"marketRisk\"].rolling(7).mean()\n",
        "\n",
        "#Drop NA\n",
        "Ana = Ana.dropna(0)\n",
        "Ana.head()"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "5448\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Close</th>\n",
              "      <th>marketRisk</th>\n",
              "      <th>returns</th>\n",
              "      <th>Target</th>\n",
              "      <th>Change</th>\n",
              "      <th>Signal</th>\n",
              "      <th>Close30</th>\n",
              "      <th>Close100</th>\n",
              "      <th>r1</th>\n",
              "      <th>r2</th>\n",
              "      <th>r3</th>\n",
              "      <th>r4</th>\n",
              "      <th>marketrisk_avg30</th>\n",
              "      <th>marketrisk_avg90</th>\n",
              "      <th>marketrisk_avg7</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Date</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1998-05-21</th>\n",
              "      <td>0.894694</td>\n",
              "      <td>0.017921</td>\n",
              "      <td>-0.005816</td>\n",
              "      <td>-0.001439</td>\n",
              "      <td>-0.001608</td>\n",
              "      <td>False</td>\n",
              "      <td>0.905687</td>\n",
              "      <td>0.916241</td>\n",
              "      <td>-0.005234</td>\n",
              "      <td>-0.007833</td>\n",
              "      <td>-0.025692</td>\n",
              "      <td>-0.031832</td>\n",
              "      <td>0.030184</td>\n",
              "      <td>0.034995</td>\n",
              "      <td>0.031504</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-22</th>\n",
              "      <td>0.893256</td>\n",
              "      <td>0.054887</td>\n",
              "      <td>-0.001608</td>\n",
              "      <td>0.004813</td>\n",
              "      <td>0.005388</td>\n",
              "      <td>True</td>\n",
              "      <td>0.904816</td>\n",
              "      <td>0.916045</td>\n",
              "      <td>-0.001439</td>\n",
              "      <td>-0.011721</td>\n",
              "      <td>-0.026115</td>\n",
              "      <td>-0.034474</td>\n",
              "      <td>0.030500</td>\n",
              "      <td>0.035283</td>\n",
              "      <td>0.035738</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-25</th>\n",
              "      <td>0.898069</td>\n",
              "      <td>0.072076</td>\n",
              "      <td>0.005388</td>\n",
              "      <td>0.001373</td>\n",
              "      <td>0.001529</td>\n",
              "      <td>True</td>\n",
              "      <td>0.904089</td>\n",
              "      <td>0.915792</td>\n",
              "      <td>0.004813</td>\n",
              "      <td>-0.005518</td>\n",
              "      <td>-0.021809</td>\n",
              "      <td>-0.030695</td>\n",
              "      <td>0.031202</td>\n",
              "      <td>0.035543</td>\n",
              "      <td>0.042474</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-26</th>\n",
              "      <td>0.899442</td>\n",
              "      <td>0.120229</td>\n",
              "      <td>0.001529</td>\n",
              "      <td>0.006027</td>\n",
              "      <td>0.006700</td>\n",
              "      <td>True</td>\n",
              "      <td>0.903746</td>\n",
              "      <td>0.915530</td>\n",
              "      <td>0.001373</td>\n",
              "      <td>-0.007587</td>\n",
              "      <td>-0.010310</td>\n",
              "      <td>-0.030358</td>\n",
              "      <td>0.034618</td>\n",
              "      <td>0.036311</td>\n",
              "      <td>0.056034</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-27</th>\n",
              "      <td>0.905469</td>\n",
              "      <td>0.057863</td>\n",
              "      <td>0.006700</td>\n",
              "      <td>-0.001718</td>\n",
              "      <td>-0.001898</td>\n",
              "      <td>False</td>\n",
              "      <td>0.903617</td>\n",
              "      <td>0.915352</td>\n",
              "      <td>0.006027</td>\n",
              "      <td>-0.001643</td>\n",
              "      <td>-0.003870</td>\n",
              "      <td>-0.016785</td>\n",
              "      <td>0.036330</td>\n",
              "      <td>0.036533</td>\n",
              "      <td>0.055463</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "               Close  marketRisk   returns    Target    Change  Signal  \\\n",
              "Date                                                                     \n",
              "1998-05-21  0.894694    0.017921 -0.005816 -0.001439 -0.001608   False   \n",
              "1998-05-22  0.893256    0.054887 -0.001608  0.004813  0.005388    True   \n",
              "1998-05-25  0.898069    0.072076  0.005388  0.001373  0.001529    True   \n",
              "1998-05-26  0.899442    0.120229  0.001529  0.006027  0.006700    True   \n",
              "1998-05-27  0.905469    0.057863  0.006700 -0.001718 -0.001898   False   \n",
              "\n",
              "             Close30  Close100        r1        r2        r3        r4  \\\n",
              "Date                                                                     \n",
              "1998-05-21  0.905687  0.916241 -0.005234 -0.007833 -0.025692 -0.031832   \n",
              "1998-05-22  0.904816  0.916045 -0.001439 -0.011721 -0.026115 -0.034474   \n",
              "1998-05-25  0.904089  0.915792  0.004813 -0.005518 -0.021809 -0.030695   \n",
              "1998-05-26  0.903746  0.915530  0.001373 -0.007587 -0.010310 -0.030358   \n",
              "1998-05-27  0.903617  0.915352  0.006027 -0.001643 -0.003870 -0.016785   \n",
              "\n",
              "            marketrisk_avg30  marketrisk_avg90  marketrisk_avg7  \n",
              "Date                                                             \n",
              "1998-05-21          0.030184          0.034995         0.031504  \n",
              "1998-05-22          0.030500          0.035283         0.035738  \n",
              "1998-05-25          0.031202          0.035543         0.042474  \n",
              "1998-05-26          0.034618          0.036311         0.056034  \n",
              "1998-05-27          0.036330          0.036533         0.055463  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "HInEXnin29AV",
        "outputId": "e7b81682-df83-4bfb-eda0-9ead06d6e07d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "cell_type": "code",
      "source": [
        "#Define function to convert dataframe to training, validating and testing examples\n",
        "def series_to_supervised(df, n_in=1, n_out=1):\n",
        "\tn_vars = 1 if type(df) is list else df.shape[1]\n",
        "  \n",
        "  \n",
        "\tcols, names = list(), list()\n",
        "\t# input sequence (t-n, ... t-1)\n",
        "\tfor i in range(n_in, -1, -1):\n",
        "\t\tcols.append(df.shift(i))\n",
        "\t\tnames += [(df.columns[j]+'(t-%d)' % (i)) for j in range(n_vars)]\n",
        "\n",
        "\t# put it all together\n",
        "\tagg = concat(cols, axis=1)\n",
        "\tagg.columns = names\n",
        "\treturn agg\n",
        "\n",
        "#Features Selection\n",
        "Set =  Ana[['r1', 'r2', 'r3','r4', 'marketRisk', 'marketrisk_avg7', 'marketrisk_avg30', 'marketrisk_avg90']]\n",
        "# Set =  Ana[['r1', 'r2']]\n",
        "FEATURES_SHAPE =  Set.shape[1]\n",
        "print(FEATURES_SHAPE)\n",
        "\n",
        "#Forming examples\n",
        "SEQ_LEN = 5\n",
        "Set =  series_to_supervised(Set, SEQ_LEN-1, 0)\n",
        "Set[[\"Change\", \"Signal\"]] = Ana[[\"Change\", \"Signal\"]]\n",
        "Set = Set.dropna()\n",
        "print(Set.shape)\n",
        "# Set.head()"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "8\n",
            "(5345, 42)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "9ar94vy_ip_7",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "#Remove bad training and validating examples\n",
        "SetAdjusted = Set.copy()\n",
        "# SetAdjusted = Set.loc[(Set.Change< -0.0015)| (Set.Change> 0.003)]\n",
        "# SetAdjusted.head()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "colab_type": "text",
        "id": "Ovn_3v_G1Xuq"
      },
      "cell_type": "markdown",
      "source": [
        "## 3.2 Train-Validate-Test Split"
      ]
    },
    {
      "metadata": {
        "id": "-1pSODY0CUY4",
        "colab_type": "code",
        "outputId": "61a97ee3-6478-42aa-d9ca-233541a19ae6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 88
        }
      },
      "cell_type": "code",
      "source": [
        "daterange = pd.date_range(start='18/1/2017', periods=2, freq='2W')\n",
        "\n",
        "print(daterange)\n",
        "print(daterange[0].strftime(\"%Y-%m-%d\"))\n",
        "len(daterange)\n",
        "\n",
        "for single_date in daterange:\n",
        "  print(single_date - relativedelta(years=1))"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "DatetimeIndex(['2017-01-22', '2017-02-05'], dtype='datetime64[ns]', freq='2W-SUN')\n",
            "2017-01-22\n",
            "2016-01-22 00:00:00\n",
            "2016-02-05 00:00:00\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "7cwIwS1Pe1Uq",
        "colab_type": "code",
        "outputId": "c761c21c-db5c-4a57-9a18-0b1ababca157",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "testyhat = []\n",
        "testyhat += [[1,2,3]]\n",
        "testyhat += [[1,2,3]]\n",
        "testyhat\n",
        "np.mean(testyhat, axis =0) > 1"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([False,  True,  True])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "JvOKmcgM1da_",
        "outputId": "bf3bb134-f16c-4485-ebe2-2b4ec5196d5d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 3444
        }
      },
      "cell_type": "code",
      "source": [
        "daterange = pd.date_range(start='1/1/2017', periods=1 , freq='4W')\n",
        "print(daterange)\n",
        "tradeTable= Ana.copy()\n",
        "tradeTable[\"predictSignal\"] = 0\n",
        "# tradeTable.head()\n",
        "EFF_TEST_START = \"\"\n",
        "EFF_TEST_END = \"\"\n",
        "num = 0 \n",
        "ENSEMBLE = 1\n",
        "\n",
        "for single_date in daterange:\n",
        "  \n",
        "  testyhat = []\n",
        "  num += 1\n",
        "      \n",
        "  for i in range(0, ENSEMBLE):\n",
        "    TRAIN_START = (single_date - relativedelta(years = 10)).strftime(\"%Y-%m-%d\")\n",
        "    TRAIN_END = (single_date - relativedelta(years =2) - timedelta(days=1)).strftime(\"%Y-%m-%d\") \n",
        "    \n",
        "    VAL_START = (single_date - relativedelta(years = 2)).strftime(\"%Y-%m-%d\")\n",
        "    VAL_END = (single_date - timedelta(days = 1)).strftime(\"%Y-%m-%d\")\n",
        "    \n",
        "    TEST_START = (single_date).strftime(\"%Y-%m-%d\")\n",
        "    TEST_END = (single_date + relativedelta(years=2)).strftime(\"%Y-%m-%d\")\n",
        "    print(TRAIN_START)\n",
        "    print(TRAIN_END)\n",
        "    print(VAL_START)\n",
        "    print(VAL_END)\n",
        "    print(TEST_START)\n",
        "    print(TEST_END)\n",
        "  \n",
        "    if num == 1:\n",
        "      EFF_TEST_START = TEST_START\n",
        "    if num == len(daterange):\n",
        "      EFF_TEST_END = TEST_END\n",
        "    \n",
        "\n",
        "    #Set x and y variables\n",
        "    FIRSTTRAINVAR = SetAdjusted.columns[0]\n",
        "    print(FIRSTTRAINVAR)\n",
        "    LASTTRAINVARIABLE = SetAdjusted.columns[-3]\n",
        "    print(LASTTRAINVARIABLE)\n",
        "    YVAR= \"Signal\"\n",
        "\n",
        "  #   #Split train and test set at TRAIN_DATE_END and TEST_DATE_START\n",
        "  #   TRAIN_START = \"2017-01-01\"\n",
        "  #   TRAIN_END = \"2017-03-31\"\n",
        "  #   # VALIDATE_START = \"2018-01-01\"\n",
        "  #   # VALIDATE_END = \"2018-05-31\"\n",
        "  #   TEST_START = \"2017-01-01\"\n",
        "  #   TEST_END = \"2017-04-30\"\n",
        "\n",
        "    #Specify LSTM input shape\n",
        "    X_SHAPE1 = 5\n",
        "    X_SHAPE2 = FEATURES_SHAPE\n",
        "\n",
        "    #train_x\n",
        "    train_x = SetAdjusted.loc[TRAIN_START:TRAIN_END, FIRSTTRAINVAR:LASTTRAINVARIABLE]\n",
        "    # Normalised train_x to [0,1]\n",
        "    # print(train_x.head())\n",
        "    scaler = MinMaxScaler(feature_range=(0, 1))\n",
        "    train_x = scaler.fit_transform(train_x)\n",
        "  #   print(train_x.shape)\n",
        "    #reshape train_x for RNN input\n",
        "    train_x = train_x.reshape(train_x.shape[0], X_SHAPE1, X_SHAPE2)\n",
        "  #   print(train_x.shape)\n",
        "\n",
        "    #train_y\n",
        "    train_y = SetAdjusted.loc[TRAIN_START:TRAIN_END, YVAR]\n",
        "    train_y = train_y.values.reshape((train_y.shape[0], 1))\n",
        "    yscaler = MinMaxScaler(feature_range=(0, 1))\n",
        "    # train_y = yscaler.fit_transform(train_y)\n",
        "  #   print(train_y.shape)\n",
        "  \n",
        "    #val_x\n",
        "    val_x = Set.loc[VAL_START:VAL_END, FIRSTTRAINVAR:LASTTRAINVARIABLE]\n",
        "    # Normalised val_x to [0,1]\n",
        "    val_x = scaler.transform(val_x)\n",
        "    #reshape val_x for RNN input\n",
        "    val_x = val_x.reshape(val_x.shape[0], X_SHAPE1, X_SHAPE2)\n",
        "  #   print(test_x.shape)\n",
        "\n",
        "    #val_y\n",
        "    val_y = Set.loc[VAL_START:VAL_END, YVAR]\n",
        "    val_y = val_y.values.reshape((val_y.shape[0], 1))\n",
        "    # test_y = yscaler.transform(test_y)\n",
        "\n",
        "    #test_x\n",
        "    test_x = Set.loc[TEST_START:TEST_END, FIRSTTRAINVAR:LASTTRAINVARIABLE]\n",
        "    # Normalised train_x to [0,1]\n",
        "    test_x = scaler.transform(test_x)\n",
        "    #reshape test_x for RNN input\n",
        "    test_x = test_x.reshape(test_x.shape[0], X_SHAPE1, X_SHAPE2)\n",
        "  #   print(test_x.shape)\n",
        "\n",
        "    #test_y\n",
        "    test_y = Set.loc[TEST_START:TEST_END, YVAR]\n",
        "    test_y = test_y.values.reshape((test_y.shape[0], 1))\n",
        "    # test_y = yscaler.transform(test_y)\n",
        "\n",
        "  #   print(test_y.shape)\n",
        "\n",
        "\n",
        "    TESTEND = test_x.shape[0]\n",
        "    # TESTSTART = train_x.shape[0] + validate_x.shape[0]\n",
        "    TESTSTART = train_x.shape[0]\n",
        "    # VALIDATESTART = train_x.shape[0]\n",
        "    print(TESTSTART)\n",
        "\n",
        "    ## Specifying Batch Size\n",
        "    n_batch = 2\n",
        "\n",
        "    ## Only to use when removing model \n",
        "    !rm best.h5\n",
        "\n",
        "    #Check train_y statistic\n",
        "    trainStats = np.unique(train_y, return_counts=True)\n",
        "    print(trainStats)\n",
        "\n",
        "    #Set Classweight\n",
        "    if trainStats[1][0] > trainStats[1][1]: \n",
        "      classweight = {0:1, 1:trainStats[1][0]/trainStats[1][1]}\n",
        "    else:\n",
        "      classweight = {0:trainStats[1][1]/trainStats[1][0], 1:1}\n",
        "\n",
        "    print(classweight)\n",
        "\n",
        "    units = 20\n",
        "    inputs = Input(shape=(train_x.shape[1], train_x.shape[2]))\n",
        "    secondary = LSTM(units, return_sequences=True)(inputs)\n",
        "    activations = LSTM(units, return_sequences=True)(secondary)\n",
        "    # out = Dense(1, activation='sigmoid')(activations)\n",
        "\n",
        "    attention = Dense(1, activation='tanh')(activations)\n",
        "    attention = Flatten()(attention)\n",
        "    attention = Activation('softmax')(attention)\n",
        "    attention = RepeatVector(units)(attention)\n",
        "    attention = Permute([2, 1])(attention)\n",
        "\n",
        "\n",
        "    sent_representation = multiply([activations, attention])\n",
        "    sent_representation = Lambda(lambda xin: K.sum(xin, axis=-2), output_shape=(units,))(sent_representation)\n",
        "\n",
        "    probabilities = Dense(1, activation='sigmoid')(sent_representation)\n",
        "\n",
        "\n",
        "    model = Model(inputs=inputs, outputs=probabilities)\n",
        "    model.compile(optimizer='rmsprop',\n",
        "                  loss='binary_crossentropy',\n",
        "                  metrics=[\"binary_accuracy\"])\n",
        "\n",
        "    # Checkpoint\n",
        "    checkpoint = ModelCheckpoint(\"best.h5\", monitor='val_binary_accuracy', verbose=1, save_best_only=True, save_weights_only=False, mode='max')\n",
        "    early = EarlyStopping(monitor='val_binary_accuracy', min_delta=0, patience=40, verbose=0, mode='auto')\n",
        "    callbacks_list = [checkpoint]\n",
        "\n",
        "    # Fit model\n",
        "    if os.path.isfile(\"best.h5\"):\n",
        "      model.load_weights(\"best.h5\")\n",
        "\n",
        "    history = model.fit(train_x, train_y, epochs=3,  batch_size=n_batch, callbacks=[checkpoint, early],\n",
        "                        validation_data=(val_x, val_y), \n",
        "                        shuffle=False, verbose =  1,\n",
        "                          class_weight= classweight\n",
        "                          )\n",
        "\n",
        "    # Make predictions, yhat, on test set\n",
        "    if os.path.isfile(\"best.h5\"):\n",
        "      model.load_weights(\"best.h5\")\n",
        "    yhat = model.predict(test_x, batch_size=n_batch)\n",
        "    # yhat = yscaler.inverse_transform(yhat)\n",
        "    print(test_x.shape)\n",
        "    print(yhat.shape)\n",
        "\n",
        "    ## Reading yhat as predictSignal\n",
        "    testyhat += [yhat]\n",
        "    print(len(testyhat))\n",
        "    \n",
        "    print(tradeTable.loc[TEST_START:TEST_END, \"predictSignal\"].shape)\n",
        "    tradeTable.loc[TEST_START:TEST_END, \"predictSignal\"] = np.mean(testyhat, axis =0) >0.5\n",
        "    \n",
        "    # tradeTable[\"predictPrice\"] = yhat\n",
        "    # tradeTable[\"predictSignal\"]= tradeTable[\"predictPrice\"]>0\n",
        "  #   tradeTable.head()\n",
        "  \n",
        "  \n",
        "\n",
        "#Changing Signals to Weights\n",
        "tradeTable['predictedweight'] = np.where(tradeTable['predictSignal'] == True, 1, -1)\n",
        "tradeTable['labelledweight'] = np.where(tradeTable['Signal'] == True, 1, -1)\n",
        "tradeTable.head()\n",
        "\n",
        "# calculating r, %tage change of closing price over previous time step\n",
        "tradeTable['r'] = (tradeTable['Close']- tradeTable['Close'].shift(1))/tradeTable['Close'].shift(1)\n",
        "tradeTable.head()\n",
        "\n",
        "# calculating predicted and labelled trade returns over each time step\n",
        "tradeTable['predictedreturns'] = tradeTable['r']*tradeTable['predictedweight'].shift(1)\n",
        "tradeTable['labelledreturns'] = tradeTable['r']*tradeTable['labelledweight'].shift(1)\n",
        "\n",
        "tradeTable"
      ],
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "DatetimeIndex(['2017-01-01'], dtype='datetime64[ns]', freq='4W-SUN')\n",
            "2007-01-01\n",
            "2014-12-31\n",
            "2015-01-01\n",
            "2016-12-31\n",
            "2017-01-01\n",
            "2019-01-01\n",
            "r1(t-4)\n",
            "marketrisk_avg90(t-0)\n",
            "2088\n",
            "(array([False,  True]), array([1073, 1015]))\n",
            "{0: 1, 1: 1.0571428571428572}\n",
            "Train on 2088 samples, validate on 522 samples\n",
            "Epoch 1/3\n",
            "2088/2088 [==============================] - 50s 24ms/step - loss: 0.7136 - binary_accuracy: 0.5043 - val_loss: 0.6926 - val_binary_accuracy: 0.5192\n",
            "\n",
            "Epoch 00001: val_binary_accuracy improved from -inf to 0.51916, saving model to best.h5\n",
            "Epoch 2/3\n",
            "2088/2088 [==============================] - 33s 16ms/step - loss: 0.7134 - binary_accuracy: 0.5000 - val_loss: 0.6926 - val_binary_accuracy: 0.5192\n",
            "\n",
            "Epoch 00002: val_binary_accuracy did not improve from 0.51916\n",
            "Epoch 3/3\n",
            "2088/2088 [==============================] - 33s 16ms/step - loss: 0.7132 - binary_accuracy: 0.4990 - val_loss: 0.6926 - val_binary_accuracy: 0.5192\n",
            "\n",
            "Epoch 00003: val_binary_accuracy did not improve from 0.51916\n",
            "(495, 5, 8)\n",
            "(495, 1)\n",
            "1\n",
            "(495,)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Close</th>\n",
              "      <th>marketRisk</th>\n",
              "      <th>returns</th>\n",
              "      <th>Target</th>\n",
              "      <th>Change</th>\n",
              "      <th>Signal</th>\n",
              "      <th>Close30</th>\n",
              "      <th>Close100</th>\n",
              "      <th>r1</th>\n",
              "      <th>r2</th>\n",
              "      <th>...</th>\n",
              "      <th>r4</th>\n",
              "      <th>marketrisk_avg30</th>\n",
              "      <th>marketrisk_avg90</th>\n",
              "      <th>marketrisk_avg7</th>\n",
              "      <th>predictSignal</th>\n",
              "      <th>predictedweight</th>\n",
              "      <th>labelledweight</th>\n",
              "      <th>r</th>\n",
              "      <th>predictedreturns</th>\n",
              "      <th>labelledreturns</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Date</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1998-05-21</th>\n",
              "      <td>0.894694</td>\n",
              "      <td>0.017921</td>\n",
              "      <td>-0.005816</td>\n",
              "      <td>-0.001439</td>\n",
              "      <td>-0.001608</td>\n",
              "      <td>False</td>\n",
              "      <td>0.905687</td>\n",
              "      <td>0.916241</td>\n",
              "      <td>-0.005234</td>\n",
              "      <td>-0.007833</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.031832</td>\n",
              "      <td>0.030184</td>\n",
              "      <td>0.034995</td>\n",
              "      <td>0.031504</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-22</th>\n",
              "      <td>0.893256</td>\n",
              "      <td>0.054887</td>\n",
              "      <td>-0.001608</td>\n",
              "      <td>0.004813</td>\n",
              "      <td>0.005388</td>\n",
              "      <td>True</td>\n",
              "      <td>0.904816</td>\n",
              "      <td>0.916045</td>\n",
              "      <td>-0.001439</td>\n",
              "      <td>-0.011721</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.034474</td>\n",
              "      <td>0.030500</td>\n",
              "      <td>0.035283</td>\n",
              "      <td>0.035738</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.001608</td>\n",
              "      <td>0.001608</td>\n",
              "      <td>0.001608</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-25</th>\n",
              "      <td>0.898069</td>\n",
              "      <td>0.072076</td>\n",
              "      <td>0.005388</td>\n",
              "      <td>0.001373</td>\n",
              "      <td>0.001529</td>\n",
              "      <td>True</td>\n",
              "      <td>0.904089</td>\n",
              "      <td>0.915792</td>\n",
              "      <td>0.004813</td>\n",
              "      <td>-0.005518</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.030695</td>\n",
              "      <td>0.031202</td>\n",
              "      <td>0.035543</td>\n",
              "      <td>0.042474</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.005388</td>\n",
              "      <td>-0.005388</td>\n",
              "      <td>0.005388</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-26</th>\n",
              "      <td>0.899442</td>\n",
              "      <td>0.120229</td>\n",
              "      <td>0.001529</td>\n",
              "      <td>0.006027</td>\n",
              "      <td>0.006700</td>\n",
              "      <td>True</td>\n",
              "      <td>0.903746</td>\n",
              "      <td>0.915530</td>\n",
              "      <td>0.001373</td>\n",
              "      <td>-0.007587</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.030358</td>\n",
              "      <td>0.034618</td>\n",
              "      <td>0.036311</td>\n",
              "      <td>0.056034</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.001529</td>\n",
              "      <td>-0.001529</td>\n",
              "      <td>0.001529</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-27</th>\n",
              "      <td>0.905469</td>\n",
              "      <td>0.057863</td>\n",
              "      <td>0.006700</td>\n",
              "      <td>-0.001718</td>\n",
              "      <td>-0.001898</td>\n",
              "      <td>False</td>\n",
              "      <td>0.903617</td>\n",
              "      <td>0.915352</td>\n",
              "      <td>0.006027</td>\n",
              "      <td>-0.001643</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.016785</td>\n",
              "      <td>0.036330</td>\n",
              "      <td>0.036533</td>\n",
              "      <td>0.055463</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.006700</td>\n",
              "      <td>-0.006700</td>\n",
              "      <td>0.006700</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-28</th>\n",
              "      <td>0.903751</td>\n",
              "      <td>0.039126</td>\n",
              "      <td>-0.001898</td>\n",
              "      <td>0.003361</td>\n",
              "      <td>0.003719</td>\n",
              "      <td>True</td>\n",
              "      <td>0.903403</td>\n",
              "      <td>0.915163</td>\n",
              "      <td>-0.001718</td>\n",
              "      <td>-0.001882</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.011916</td>\n",
              "      <td>0.035828</td>\n",
              "      <td>0.036895</td>\n",
              "      <td>0.055423</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.001898</td>\n",
              "      <td>0.001898</td>\n",
              "      <td>0.001898</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-05-29</th>\n",
              "      <td>0.907112</td>\n",
              "      <td>0.044906</td>\n",
              "      <td>0.003719</td>\n",
              "      <td>-0.004096</td>\n",
              "      <td>-0.004515</td>\n",
              "      <td>False</td>\n",
              "      <td>0.903251</td>\n",
              "      <td>0.915000</td>\n",
              "      <td>0.003361</td>\n",
              "      <td>0.007184</td>\n",
              "      <td>...</td>\n",
              "      <td>0.004422</td>\n",
              "      <td>0.036237</td>\n",
              "      <td>0.037338</td>\n",
              "      <td>0.058144</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.003719</td>\n",
              "      <td>-0.003719</td>\n",
              "      <td>0.003719</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-01</th>\n",
              "      <td>0.903016</td>\n",
              "      <td>0.082701</td>\n",
              "      <td>-0.004515</td>\n",
              "      <td>0.000734</td>\n",
              "      <td>0.000813</td>\n",
              "      <td>True</td>\n",
              "      <td>0.903043</td>\n",
              "      <td>0.914801</td>\n",
              "      <td>-0.004096</td>\n",
              "      <td>0.008322</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.007233</td>\n",
              "      <td>0.038099</td>\n",
              "      <td>0.037892</td>\n",
              "      <td>0.067398</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.004515</td>\n",
              "      <td>0.004515</td>\n",
              "      <td>0.004515</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-02</th>\n",
              "      <td>0.903751</td>\n",
              "      <td>0.031756</td>\n",
              "      <td>0.000813</td>\n",
              "      <td>-0.005278</td>\n",
              "      <td>-0.005840</td>\n",
              "      <td>False</td>\n",
              "      <td>0.902986</td>\n",
              "      <td>0.914626</td>\n",
              "      <td>0.000734</td>\n",
              "      <td>0.010495</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.003938</td>\n",
              "      <td>0.038561</td>\n",
              "      <td>0.038163</td>\n",
              "      <td>0.064094</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.000813</td>\n",
              "      <td>-0.000813</td>\n",
              "      <td>0.000813</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-03</th>\n",
              "      <td>0.898473</td>\n",
              "      <td>0.005857</td>\n",
              "      <td>-0.005840</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>False</td>\n",
              "      <td>0.902681</td>\n",
              "      <td>0.914384</td>\n",
              "      <td>-0.005278</td>\n",
              "      <td>0.000403</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.021321</td>\n",
              "      <td>0.038295</td>\n",
              "      <td>0.037822</td>\n",
              "      <td>0.054634</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.005840</td>\n",
              "      <td>0.005840</td>\n",
              "      <td>0.005840</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-04</th>\n",
              "      <td>0.898473</td>\n",
              "      <td>0.020549</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.002428</td>\n",
              "      <td>0.002703</td>\n",
              "      <td>True</td>\n",
              "      <td>0.902308</td>\n",
              "      <td>0.914103</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-0.000970</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.027968</td>\n",
              "      <td>0.037835</td>\n",
              "      <td>0.037518</td>\n",
              "      <td>0.040394</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-0.000000</td>\n",
              "      <td>-0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-05</th>\n",
              "      <td>0.900901</td>\n",
              "      <td>0.033363</td>\n",
              "      <td>0.002703</td>\n",
              "      <td>0.002441</td>\n",
              "      <td>0.002710</td>\n",
              "      <td>True</td>\n",
              "      <td>0.902107</td>\n",
              "      <td>0.913835</td>\n",
              "      <td>0.002428</td>\n",
              "      <td>-0.004568</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.027173</td>\n",
              "      <td>0.038003</td>\n",
              "      <td>0.037519</td>\n",
              "      <td>0.036894</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.002703</td>\n",
              "      <td>-0.002703</td>\n",
              "      <td>0.002703</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-08</th>\n",
              "      <td>0.903342</td>\n",
              "      <td>0.055741</td>\n",
              "      <td>0.002710</td>\n",
              "      <td>-0.000245</td>\n",
              "      <td>-0.000271</td>\n",
              "      <td>False</td>\n",
              "      <td>0.902044</td>\n",
              "      <td>0.913581</td>\n",
              "      <td>0.002441</td>\n",
              "      <td>-0.000408</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.017892</td>\n",
              "      <td>0.037301</td>\n",
              "      <td>0.037236</td>\n",
              "      <td>0.039268</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.002710</td>\n",
              "      <td>-0.002710</td>\n",
              "      <td>0.002710</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-09</th>\n",
              "      <td>0.903098</td>\n",
              "      <td>0.015388</td>\n",
              "      <td>-0.000271</td>\n",
              "      <td>0.008064</td>\n",
              "      <td>0.008929</td>\n",
              "      <td>True</td>\n",
              "      <td>0.901880</td>\n",
              "      <td>0.913314</td>\n",
              "      <td>-0.000245</td>\n",
              "      <td>-0.004014</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.015345</td>\n",
              "      <td>0.036329</td>\n",
              "      <td>0.036582</td>\n",
              "      <td>0.035051</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.000271</td>\n",
              "      <td>0.000271</td>\n",
              "      <td>0.000271</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-10</th>\n",
              "      <td>0.911162</td>\n",
              "      <td>0.049067</td>\n",
              "      <td>0.008929</td>\n",
              "      <td>0.006775</td>\n",
              "      <td>0.007435</td>\n",
              "      <td>True</td>\n",
              "      <td>0.901968</td>\n",
              "      <td>0.913203</td>\n",
              "      <td>0.008064</td>\n",
              "      <td>0.008146</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.002080</td>\n",
              "      <td>0.036190</td>\n",
              "      <td>0.036510</td>\n",
              "      <td>0.030246</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.008929</td>\n",
              "      <td>-0.008929</td>\n",
              "      <td>0.008929</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-11</th>\n",
              "      <td>0.917936</td>\n",
              "      <td>0.067583</td>\n",
              "      <td>0.007435</td>\n",
              "      <td>-0.003191</td>\n",
              "      <td>-0.003476</td>\n",
              "      <td>False</td>\n",
              "      <td>0.902304</td>\n",
              "      <td>0.913226</td>\n",
              "      <td>0.006775</td>\n",
              "      <td>0.014186</td>\n",
              "      <td>...</td>\n",
              "      <td>0.011072</td>\n",
              "      <td>0.037385</td>\n",
              "      <td>0.036613</td>\n",
              "      <td>0.035364</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.007435</td>\n",
              "      <td>-0.007435</td>\n",
              "      <td>0.007435</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-12</th>\n",
              "      <td>0.914746</td>\n",
              "      <td>0.057471</td>\n",
              "      <td>-0.003476</td>\n",
              "      <td>0.002013</td>\n",
              "      <td>0.002200</td>\n",
              "      <td>True</td>\n",
              "      <td>0.902717</td>\n",
              "      <td>0.913346</td>\n",
              "      <td>-0.003191</td>\n",
              "      <td>0.016273</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.002854</td>\n",
              "      <td>0.038482</td>\n",
              "      <td>0.037026</td>\n",
              "      <td>0.042737</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.003476</td>\n",
              "      <td>0.003476</td>\n",
              "      <td>0.003476</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-15</th>\n",
              "      <td>0.916758</td>\n",
              "      <td>0.058427</td>\n",
              "      <td>0.002200</td>\n",
              "      <td>-0.006012</td>\n",
              "      <td>-0.006557</td>\n",
              "      <td>False</td>\n",
              "      <td>0.903248</td>\n",
              "      <td>0.913411</td>\n",
              "      <td>0.002013</td>\n",
              "      <td>0.018286</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.005666</td>\n",
              "      <td>0.039980</td>\n",
              "      <td>0.037746</td>\n",
              "      <td>0.048149</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.002200</td>\n",
              "      <td>-0.002200</td>\n",
              "      <td>0.002200</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-16</th>\n",
              "      <td>0.910747</td>\n",
              "      <td>0.017422</td>\n",
              "      <td>-0.006557</td>\n",
              "      <td>-0.006506</td>\n",
              "      <td>-0.007144</td>\n",
              "      <td>False</td>\n",
              "      <td>0.903711</td>\n",
              "      <td>0.913442</td>\n",
              "      <td>-0.006012</td>\n",
              "      <td>0.009846</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.006853</td>\n",
              "      <td>0.040951</td>\n",
              "      <td>0.037204</td>\n",
              "      <td>0.045871</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.006557</td>\n",
              "      <td>0.006557</td>\n",
              "      <td>0.006557</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-17</th>\n",
              "      <td>0.904241</td>\n",
              "      <td>0.018393</td>\n",
              "      <td>-0.007144</td>\n",
              "      <td>0.002460</td>\n",
              "      <td>0.002720</td>\n",
              "      <td>True</td>\n",
              "      <td>0.903987</td>\n",
              "      <td>0.913286</td>\n",
              "      <td>-0.006506</td>\n",
              "      <td>0.000899</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.017418</td>\n",
              "      <td>0.040985</td>\n",
              "      <td>0.037192</td>\n",
              "      <td>0.040536</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.007144</td>\n",
              "      <td>0.007144</td>\n",
              "      <td>0.007144</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-18</th>\n",
              "      <td>0.906701</td>\n",
              "      <td>0.054887</td>\n",
              "      <td>0.002720</td>\n",
              "      <td>-0.002296</td>\n",
              "      <td>-0.002532</td>\n",
              "      <td>False</td>\n",
              "      <td>0.904282</td>\n",
              "      <td>0.913089</td>\n",
              "      <td>0.002460</td>\n",
              "      <td>0.003603</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.009470</td>\n",
              "      <td>0.042129</td>\n",
              "      <td>0.037509</td>\n",
              "      <td>0.046179</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.002720</td>\n",
              "      <td>-0.002720</td>\n",
              "      <td>0.002720</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-19</th>\n",
              "      <td>0.904404</td>\n",
              "      <td>0.027351</td>\n",
              "      <td>-0.002532</td>\n",
              "      <td>0.003037</td>\n",
              "      <td>0.003358</td>\n",
              "      <td>True</td>\n",
              "      <td>0.904448</td>\n",
              "      <td>0.912852</td>\n",
              "      <td>-0.002296</td>\n",
              "      <td>-0.006757</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.016236</td>\n",
              "      <td>0.041939</td>\n",
              "      <td>0.037346</td>\n",
              "      <td>0.043076</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.002532</td>\n",
              "      <td>0.002532</td>\n",
              "      <td>0.002532</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-22</th>\n",
              "      <td>0.907441</td>\n",
              "      <td>0.061778</td>\n",
              "      <td>0.003358</td>\n",
              "      <td>0.003389</td>\n",
              "      <td>0.003734</td>\n",
              "      <td>True</td>\n",
              "      <td>0.904593</td>\n",
              "      <td>0.912714</td>\n",
              "      <td>0.003037</td>\n",
              "      <td>-0.010495</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.011930</td>\n",
              "      <td>0.043055</td>\n",
              "      <td>0.037638</td>\n",
              "      <td>0.042247</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.003358</td>\n",
              "      <td>-0.003358</td>\n",
              "      <td>0.003358</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-23</th>\n",
              "      <td>0.910830</td>\n",
              "      <td>0.055145</td>\n",
              "      <td>0.003734</td>\n",
              "      <td>-0.000332</td>\n",
              "      <td>-0.000364</td>\n",
              "      <td>False</td>\n",
              "      <td>0.904869</td>\n",
              "      <td>0.912638</td>\n",
              "      <td>0.003389</td>\n",
              "      <td>-0.003916</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.012787</td>\n",
              "      <td>0.044083</td>\n",
              "      <td>0.037913</td>\n",
              "      <td>0.041915</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.003734</td>\n",
              "      <td>-0.003734</td>\n",
              "      <td>0.003734</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-24</th>\n",
              "      <td>0.910498</td>\n",
              "      <td>0.049160</td>\n",
              "      <td>-0.000364</td>\n",
              "      <td>-0.002150</td>\n",
              "      <td>-0.002362</td>\n",
              "      <td>False</td>\n",
              "      <td>0.905053</td>\n",
              "      <td>0.912611</td>\n",
              "      <td>-0.000332</td>\n",
              "      <td>-0.006260</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.010991</td>\n",
              "      <td>0.044880</td>\n",
              "      <td>0.038555</td>\n",
              "      <td>0.040591</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.000364</td>\n",
              "      <td>0.000364</td>\n",
              "      <td>0.000364</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-25</th>\n",
              "      <td>0.908348</td>\n",
              "      <td>0.043022</td>\n",
              "      <td>-0.002362</td>\n",
              "      <td>0.005980</td>\n",
              "      <td>0.006583</td>\n",
              "      <td>True</td>\n",
              "      <td>0.905212</td>\n",
              "      <td>0.912626</td>\n",
              "      <td>-0.002150</td>\n",
              "      <td>-0.002399</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.010517</td>\n",
              "      <td>0.045483</td>\n",
              "      <td>0.038685</td>\n",
              "      <td>0.044248</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.002362</td>\n",
              "      <td>0.002362</td>\n",
              "      <td>0.002362</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-26</th>\n",
              "      <td>0.914328</td>\n",
              "      <td>0.052612</td>\n",
              "      <td>0.006583</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>False</td>\n",
              "      <td>0.905455</td>\n",
              "      <td>0.912593</td>\n",
              "      <td>0.005980</td>\n",
              "      <td>0.010087</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.006483</td>\n",
              "      <td>0.046393</td>\n",
              "      <td>0.038958</td>\n",
              "      <td>0.049136</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.006583</td>\n",
              "      <td>-0.006583</td>\n",
              "      <td>0.006583</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-29</th>\n",
              "      <td>0.914328</td>\n",
              "      <td>0.021773</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-0.000835</td>\n",
              "      <td>-0.000913</td>\n",
              "      <td>False</td>\n",
              "      <td>0.905696</td>\n",
              "      <td>0.912512</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.007627</td>\n",
              "      <td>...</td>\n",
              "      <td>0.005154</td>\n",
              "      <td>0.045057</td>\n",
              "      <td>0.038731</td>\n",
              "      <td>0.044406</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-0.000000</td>\n",
              "      <td>-0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-06-30</th>\n",
              "      <td>0.913492</td>\n",
              "      <td>0.021075</td>\n",
              "      <td>-0.000913</td>\n",
              "      <td>0.004781</td>\n",
              "      <td>0.005234</td>\n",
              "      <td>True</td>\n",
              "      <td>0.905958</td>\n",
              "      <td>0.912471</td>\n",
              "      <td>-0.000835</td>\n",
              "      <td>0.009088</td>\n",
              "      <td>...</td>\n",
              "      <td>0.003740</td>\n",
              "      <td>0.044446</td>\n",
              "      <td>0.038792</td>\n",
              "      <td>0.043509</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.000913</td>\n",
              "      <td>0.000913</td>\n",
              "      <td>0.000913</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1998-07-01</th>\n",
              "      <td>0.918274</td>\n",
              "      <td>0.013918</td>\n",
              "      <td>0.005234</td>\n",
              "      <td>0.000928</td>\n",
              "      <td>0.001011</td>\n",
              "      <td>True</td>\n",
              "      <td>0.906569</td>\n",
              "      <td>0.912437</td>\n",
              "      <td>0.004781</td>\n",
              "      <td>0.010833</td>\n",
              "      <td>...</td>\n",
              "      <td>-0.001098</td>\n",
              "      <td>0.044048</td>\n",
              "      <td>0.038531</td>\n",
              "      <td>0.036672</td>\n",
              "      <td>0</td>\n",
              "      <td>-1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.005234</td>\n",
              "      <td>-0.005234</td>\n",
              "      <td>0.005234</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-16</th>\n",
              "      <td>0.864080</td>\n",
              "      <td>0.046064</td>\n",
              "      <td>0.000346</td>\n",
              "      <td>0.005561</td>\n",
              "      <td>0.006435</td>\n",
              "      <td>True</td>\n",
              "      <td>0.861075</td>\n",
              "      <td>0.860146</td>\n",
              "      <td>0.000299</td>\n",
              "      <td>-0.003749</td>\n",
              "      <td>...</td>\n",
              "      <td>0.012509</td>\n",
              "      <td>0.047195</td>\n",
              "      <td>0.043938</td>\n",
              "      <td>0.043278</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.000346</td>\n",
              "      <td>0.000346</td>\n",
              "      <td>0.000346</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-17</th>\n",
              "      <td>0.869641</td>\n",
              "      <td>0.063355</td>\n",
              "      <td>0.006435</td>\n",
              "      <td>0.003569</td>\n",
              "      <td>0.004104</td>\n",
              "      <td>True</td>\n",
              "      <td>0.861399</td>\n",
              "      <td>0.860267</td>\n",
              "      <td>0.005561</td>\n",
              "      <td>-0.000681</td>\n",
              "      <td>...</td>\n",
              "      <td>0.021392</td>\n",
              "      <td>0.047701</td>\n",
              "      <td>0.044340</td>\n",
              "      <td>0.044644</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.006435</td>\n",
              "      <td>0.006435</td>\n",
              "      <td>0.006435</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-18</th>\n",
              "      <td>0.873210</td>\n",
              "      <td>0.063601</td>\n",
              "      <td>0.004104</td>\n",
              "      <td>-0.004627</td>\n",
              "      <td>-0.005298</td>\n",
              "      <td>False</td>\n",
              "      <td>0.861822</td>\n",
              "      <td>0.860445</td>\n",
              "      <td>0.003569</td>\n",
              "      <td>0.002812</td>\n",
              "      <td>...</td>\n",
              "      <td>0.008682</td>\n",
              "      <td>0.049071</td>\n",
              "      <td>0.044886</td>\n",
              "      <td>0.048360</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.004104</td>\n",
              "      <td>0.004104</td>\n",
              "      <td>0.004104</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-19</th>\n",
              "      <td>0.868583</td>\n",
              "      <td>0.035059</td>\n",
              "      <td>-0.005298</td>\n",
              "      <td>0.003789</td>\n",
              "      <td>0.004362</td>\n",
              "      <td>True</td>\n",
              "      <td>0.861918</td>\n",
              "      <td>0.860554</td>\n",
              "      <td>-0.004627</td>\n",
              "      <td>0.000377</td>\n",
              "      <td>...</td>\n",
              "      <td>0.007034</td>\n",
              "      <td>0.048989</td>\n",
              "      <td>0.044557</td>\n",
              "      <td>0.046086</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.005298</td>\n",
              "      <td>-0.005298</td>\n",
              "      <td>0.005298</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-22</th>\n",
              "      <td>0.872372</td>\n",
              "      <td>0.063041</td>\n",
              "      <td>0.004362</td>\n",
              "      <td>-0.000456</td>\n",
              "      <td>-0.000523</td>\n",
              "      <td>False</td>\n",
              "      <td>0.862241</td>\n",
              "      <td>0.860729</td>\n",
              "      <td>0.003789</td>\n",
              "      <td>0.009782</td>\n",
              "      <td>...</td>\n",
              "      <td>0.011935</td>\n",
              "      <td>0.049574</td>\n",
              "      <td>0.044681</td>\n",
              "      <td>0.051670</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.004362</td>\n",
              "      <td>0.004362</td>\n",
              "      <td>0.004362</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-23</th>\n",
              "      <td>0.871916</td>\n",
              "      <td>0.024420</td>\n",
              "      <td>-0.000523</td>\n",
              "      <td>0.005970</td>\n",
              "      <td>0.006848</td>\n",
              "      <td>True</td>\n",
              "      <td>0.862582</td>\n",
              "      <td>0.860914</td>\n",
              "      <td>-0.000456</td>\n",
              "      <td>0.007013</td>\n",
              "      <td>...</td>\n",
              "      <td>0.008954</td>\n",
              "      <td>0.048726</td>\n",
              "      <td>0.044506</td>\n",
              "      <td>0.045925</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.000523</td>\n",
              "      <td>-0.000523</td>\n",
              "      <td>0.000523</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-24</th>\n",
              "      <td>0.877886</td>\n",
              "      <td>0.039890</td>\n",
              "      <td>0.006848</td>\n",
              "      <td>0.001312</td>\n",
              "      <td>0.001495</td>\n",
              "      <td>True</td>\n",
              "      <td>0.863169</td>\n",
              "      <td>0.861199</td>\n",
              "      <td>0.005970</td>\n",
              "      <td>0.014104</td>\n",
              "      <td>...</td>\n",
              "      <td>0.013582</td>\n",
              "      <td>0.049034</td>\n",
              "      <td>0.044314</td>\n",
              "      <td>0.047919</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.006848</td>\n",
              "      <td>0.006848</td>\n",
              "      <td>0.006848</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-25</th>\n",
              "      <td>0.879198</td>\n",
              "      <td>0.045768</td>\n",
              "      <td>0.001495</td>\n",
              "      <td>-0.002082</td>\n",
              "      <td>-0.002368</td>\n",
              "      <td>False</td>\n",
              "      <td>0.863956</td>\n",
              "      <td>0.861514</td>\n",
              "      <td>0.001312</td>\n",
              "      <td>0.015118</td>\n",
              "      <td>...</td>\n",
              "      <td>0.017204</td>\n",
              "      <td>0.049487</td>\n",
              "      <td>0.044216</td>\n",
              "      <td>0.047876</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.001495</td>\n",
              "      <td>0.001495</td>\n",
              "      <td>0.001495</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-26</th>\n",
              "      <td>0.877116</td>\n",
              "      <td>0.046744</td>\n",
              "      <td>-0.002368</td>\n",
              "      <td>0.002314</td>\n",
              "      <td>0.002638</td>\n",
              "      <td>True</td>\n",
              "      <td>0.864527</td>\n",
              "      <td>0.861786</td>\n",
              "      <td>-0.002082</td>\n",
              "      <td>0.007475</td>\n",
              "      <td>...</td>\n",
              "      <td>0.019115</td>\n",
              "      <td>0.049385</td>\n",
              "      <td>0.044216</td>\n",
              "      <td>0.045503</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.002368</td>\n",
              "      <td>-0.002368</td>\n",
              "      <td>0.002368</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-29</th>\n",
              "      <td>0.879430</td>\n",
              "      <td>0.060689</td>\n",
              "      <td>0.002638</td>\n",
              "      <td>0.002093</td>\n",
              "      <td>0.002380</td>\n",
              "      <td>True</td>\n",
              "      <td>0.865309</td>\n",
              "      <td>0.862093</td>\n",
              "      <td>0.002314</td>\n",
              "      <td>0.006220</td>\n",
              "      <td>...</td>\n",
              "      <td>0.024875</td>\n",
              "      <td>0.050389</td>\n",
              "      <td>0.044624</td>\n",
              "      <td>0.045087</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.002638</td>\n",
              "      <td>0.002638</td>\n",
              "      <td>0.002638</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-30</th>\n",
              "      <td>0.881523</td>\n",
              "      <td>0.037025</td>\n",
              "      <td>0.002380</td>\n",
              "      <td>0.002650</td>\n",
              "      <td>0.003006</td>\n",
              "      <td>True</td>\n",
              "      <td>0.866118</td>\n",
              "      <td>0.862392</td>\n",
              "      <td>0.002093</td>\n",
              "      <td>0.012940</td>\n",
              "      <td>...</td>\n",
              "      <td>0.022786</td>\n",
              "      <td>0.050033</td>\n",
              "      <td>0.044629</td>\n",
              "      <td>0.045368</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.002380</td>\n",
              "      <td>0.002380</td>\n",
              "      <td>0.002380</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-10-31</th>\n",
              "      <td>0.884173</td>\n",
              "      <td>0.065147</td>\n",
              "      <td>0.003006</td>\n",
              "      <td>-0.007519</td>\n",
              "      <td>-0.008504</td>\n",
              "      <td>False</td>\n",
              "      <td>0.867032</td>\n",
              "      <td>0.862752</td>\n",
              "      <td>0.002650</td>\n",
              "      <td>0.011801</td>\n",
              "      <td>...</td>\n",
              "      <td>0.018522</td>\n",
              "      <td>0.050838</td>\n",
              "      <td>0.044868</td>\n",
              "      <td>0.045669</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.003006</td>\n",
              "      <td>0.003006</td>\n",
              "      <td>0.003006</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-01</th>\n",
              "      <td>0.876655</td>\n",
              "      <td>0.038183</td>\n",
              "      <td>-0.008504</td>\n",
              "      <td>0.001694</td>\n",
              "      <td>0.001932</td>\n",
              "      <td>True</td>\n",
              "      <td>0.867945</td>\n",
              "      <td>0.862873</td>\n",
              "      <td>-0.007519</td>\n",
              "      <td>0.004739</td>\n",
              "      <td>...</td>\n",
              "      <td>0.012126</td>\n",
              "      <td>0.051000</td>\n",
              "      <td>0.044865</td>\n",
              "      <td>0.047635</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.008504</td>\n",
              "      <td>-0.008504</td>\n",
              "      <td>0.008504</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-02</th>\n",
              "      <td>0.878349</td>\n",
              "      <td>0.064924</td>\n",
              "      <td>0.001932</td>\n",
              "      <td>-0.001617</td>\n",
              "      <td>-0.001841</td>\n",
              "      <td>False</td>\n",
              "      <td>0.868852</td>\n",
              "      <td>0.863041</td>\n",
              "      <td>0.001694</td>\n",
              "      <td>0.000463</td>\n",
              "      <td>...</td>\n",
              "      <td>0.022404</td>\n",
              "      <td>0.051805</td>\n",
              "      <td>0.044856</td>\n",
              "      <td>0.051211</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.001932</td>\n",
              "      <td>0.001932</td>\n",
              "      <td>0.001932</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-05</th>\n",
              "      <td>0.876732</td>\n",
              "      <td>0.045471</td>\n",
              "      <td>-0.001841</td>\n",
              "      <td>-0.001535</td>\n",
              "      <td>-0.001750</td>\n",
              "      <td>False</td>\n",
              "      <td>0.869698</td>\n",
              "      <td>0.863204</td>\n",
              "      <td>-0.001617</td>\n",
              "      <td>-0.002467</td>\n",
              "      <td>...</td>\n",
              "      <td>0.017551</td>\n",
              "      <td>0.051697</td>\n",
              "      <td>0.044835</td>\n",
              "      <td>0.051169</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.001841</td>\n",
              "      <td>-0.001841</td>\n",
              "      <td>0.001841</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-06</th>\n",
              "      <td>0.875197</td>\n",
              "      <td>0.026117</td>\n",
              "      <td>-0.001750</td>\n",
              "      <td>0.000153</td>\n",
              "      <td>0.000175</td>\n",
              "      <td>True</td>\n",
              "      <td>0.870553</td>\n",
              "      <td>0.863326</td>\n",
              "      <td>-0.001535</td>\n",
              "      <td>-0.001919</td>\n",
              "      <td>...</td>\n",
              "      <td>0.017196</td>\n",
              "      <td>0.050886</td>\n",
              "      <td>0.044647</td>\n",
              "      <td>0.048222</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.001750</td>\n",
              "      <td>-0.001750</td>\n",
              "      <td>0.001750</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-07</th>\n",
              "      <td>0.875350</td>\n",
              "      <td>0.025385</td>\n",
              "      <td>0.000175</td>\n",
              "      <td>0.004777</td>\n",
              "      <td>0.005457</td>\n",
              "      <td>True</td>\n",
              "      <td>0.871334</td>\n",
              "      <td>0.863437</td>\n",
              "      <td>0.000153</td>\n",
              "      <td>-0.004080</td>\n",
              "      <td>...</td>\n",
              "      <td>0.017276</td>\n",
              "      <td>0.049994</td>\n",
              "      <td>0.044518</td>\n",
              "      <td>0.043179</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.000175</td>\n",
              "      <td>0.000175</td>\n",
              "      <td>0.000175</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-08</th>\n",
              "      <td>0.880127</td>\n",
              "      <td>0.050508</td>\n",
              "      <td>0.005457</td>\n",
              "      <td>0.002174</td>\n",
              "      <td>0.002470</td>\n",
              "      <td>True</td>\n",
              "      <td>0.872032</td>\n",
              "      <td>0.863618</td>\n",
              "      <td>0.004777</td>\n",
              "      <td>-0.001397</td>\n",
              "      <td>...</td>\n",
              "      <td>0.024622</td>\n",
              "      <td>0.049714</td>\n",
              "      <td>0.044651</td>\n",
              "      <td>0.045105</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.005457</td>\n",
              "      <td>0.005457</td>\n",
              "      <td>0.005457</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-09</th>\n",
              "      <td>0.882301</td>\n",
              "      <td>0.060366</td>\n",
              "      <td>0.002470</td>\n",
              "      <td>0.009203</td>\n",
              "      <td>0.010431</td>\n",
              "      <td>True</td>\n",
              "      <td>0.872726</td>\n",
              "      <td>0.863861</td>\n",
              "      <td>0.002174</td>\n",
              "      <td>-0.001872</td>\n",
              "      <td>...</td>\n",
              "      <td>0.030875</td>\n",
              "      <td>0.049158</td>\n",
              "      <td>0.044969</td>\n",
              "      <td>0.044422</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.002470</td>\n",
              "      <td>0.002470</td>\n",
              "      <td>0.002470</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-12</th>\n",
              "      <td>0.891504</td>\n",
              "      <td>0.069956</td>\n",
              "      <td>0.010431</td>\n",
              "      <td>-0.005686</td>\n",
              "      <td>-0.006378</td>\n",
              "      <td>False</td>\n",
              "      <td>0.873648</td>\n",
              "      <td>0.864230</td>\n",
              "      <td>0.009203</td>\n",
              "      <td>0.014849</td>\n",
              "      <td>...</td>\n",
              "      <td>0.040368</td>\n",
              "      <td>0.049569</td>\n",
              "      <td>0.045400</td>\n",
              "      <td>0.048961</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.010431</td>\n",
              "      <td>0.010431</td>\n",
              "      <td>0.010431</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-13</th>\n",
              "      <td>0.885818</td>\n",
              "      <td>0.045668</td>\n",
              "      <td>-0.006378</td>\n",
              "      <td>-0.001488</td>\n",
              "      <td>-0.001680</td>\n",
              "      <td>False</td>\n",
              "      <td>0.874305</td>\n",
              "      <td>0.864501</td>\n",
              "      <td>-0.005686</td>\n",
              "      <td>0.007469</td>\n",
              "      <td>...</td>\n",
              "      <td>0.034174</td>\n",
              "      <td>0.049300</td>\n",
              "      <td>0.045223</td>\n",
              "      <td>0.046210</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.006378</td>\n",
              "      <td>-0.006378</td>\n",
              "      <td>0.006378</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-14</th>\n",
              "      <td>0.884330</td>\n",
              "      <td>0.056054</td>\n",
              "      <td>-0.001680</td>\n",
              "      <td>-0.001405</td>\n",
              "      <td>-0.001589</td>\n",
              "      <td>False</td>\n",
              "      <td>0.874737</td>\n",
              "      <td>0.864688</td>\n",
              "      <td>-0.001488</td>\n",
              "      <td>0.007598</td>\n",
              "      <td>...</td>\n",
              "      <td>0.027578</td>\n",
              "      <td>0.048751</td>\n",
              "      <td>0.045235</td>\n",
              "      <td>0.047722</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.001680</td>\n",
              "      <td>-0.001680</td>\n",
              "      <td>0.001680</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-15</th>\n",
              "      <td>0.882924</td>\n",
              "      <td>0.041407</td>\n",
              "      <td>-0.001589</td>\n",
              "      <td>-0.007114</td>\n",
              "      <td>-0.008057</td>\n",
              "      <td>False</td>\n",
              "      <td>0.875215</td>\n",
              "      <td>0.864872</td>\n",
              "      <td>-0.001405</td>\n",
              "      <td>0.007727</td>\n",
              "      <td>...</td>\n",
              "      <td>0.026100</td>\n",
              "      <td>0.047424</td>\n",
              "      <td>0.044973</td>\n",
              "      <td>0.049906</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.001589</td>\n",
              "      <td>-0.001589</td>\n",
              "      <td>0.001589</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-16</th>\n",
              "      <td>0.875810</td>\n",
              "      <td>0.029247</td>\n",
              "      <td>-0.008057</td>\n",
              "      <td>-0.002524</td>\n",
              "      <td>-0.002882</td>\n",
              "      <td>False</td>\n",
              "      <td>0.875481</td>\n",
              "      <td>0.865071</td>\n",
              "      <td>-0.007114</td>\n",
              "      <td>0.000460</td>\n",
              "      <td>...</td>\n",
              "      <td>0.020012</td>\n",
              "      <td>0.046699</td>\n",
              "      <td>0.044435</td>\n",
              "      <td>0.050458</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.008057</td>\n",
              "      <td>-0.008057</td>\n",
              "      <td>0.008057</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-19</th>\n",
              "      <td>0.873286</td>\n",
              "      <td>0.006334</td>\n",
              "      <td>-0.002882</td>\n",
              "      <td>0.006221</td>\n",
              "      <td>0.007124</td>\n",
              "      <td>True</td>\n",
              "      <td>0.875580</td>\n",
              "      <td>0.865212</td>\n",
              "      <td>-0.002524</td>\n",
              "      <td>-0.006841</td>\n",
              "      <td>...</td>\n",
              "      <td>0.019242</td>\n",
              "      <td>0.045117</td>\n",
              "      <td>0.043940</td>\n",
              "      <td>0.044147</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.002882</td>\n",
              "      <td>-0.002882</td>\n",
              "      <td>0.002882</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-20</th>\n",
              "      <td>0.879507</td>\n",
              "      <td>0.033459</td>\n",
              "      <td>0.007124</td>\n",
              "      <td>-0.001004</td>\n",
              "      <td>-0.001142</td>\n",
              "      <td>False</td>\n",
              "      <td>0.875883</td>\n",
              "      <td>0.865427</td>\n",
              "      <td>0.006221</td>\n",
              "      <td>-0.002794</td>\n",
              "      <td>...</td>\n",
              "      <td>0.021801</td>\n",
              "      <td>0.044979</td>\n",
              "      <td>0.043843</td>\n",
              "      <td>0.040304</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>0.007124</td>\n",
              "      <td>0.007124</td>\n",
              "      <td>0.007124</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-21</th>\n",
              "      <td>0.878503</td>\n",
              "      <td>0.030244</td>\n",
              "      <td>-0.001142</td>\n",
              "      <td>-0.001848</td>\n",
              "      <td>-0.002104</td>\n",
              "      <td>False</td>\n",
              "      <td>0.876226</td>\n",
              "      <td>0.865631</td>\n",
              "      <td>-0.001004</td>\n",
              "      <td>-0.013001</td>\n",
              "      <td>...</td>\n",
              "      <td>0.019249</td>\n",
              "      <td>0.044288</td>\n",
              "      <td>0.043271</td>\n",
              "      <td>0.034630</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>-1</td>\n",
              "      <td>-0.001142</td>\n",
              "      <td>-0.001142</td>\n",
              "      <td>0.001142</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-22</th>\n",
              "      <td>0.876655</td>\n",
              "      <td>0.041783</td>\n",
              "      <td>-0.002104</td>\n",
              "      <td>0.005180</td>\n",
              "      <td>0.005908</td>\n",
              "      <td>True</td>\n",
              "      <td>0.876695</td>\n",
              "      <td>0.865843</td>\n",
              "      <td>-0.001848</td>\n",
              "      <td>-0.009163</td>\n",
              "      <td>...</td>\n",
              "      <td>0.017622</td>\n",
              "      <td>0.044882</td>\n",
              "      <td>0.043111</td>\n",
              "      <td>0.034075</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>-0.002104</td>\n",
              "      <td>-0.002104</td>\n",
              "      <td>0.002104</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-23</th>\n",
              "      <td>0.881834</td>\n",
              "      <td>0.031853</td>\n",
              "      <td>0.005908</td>\n",
              "      <td>0.001090</td>\n",
              "      <td>0.001236</td>\n",
              "      <td>True</td>\n",
              "      <td>0.877260</td>\n",
              "      <td>0.866147</td>\n",
              "      <td>0.005180</td>\n",
              "      <td>-0.002495</td>\n",
              "      <td>...</td>\n",
              "      <td>0.028446</td>\n",
              "      <td>0.043790</td>\n",
              "      <td>0.043037</td>\n",
              "      <td>0.030618</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.005908</td>\n",
              "      <td>0.005908</td>\n",
              "      <td>0.005908</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-11-26</th>\n",
              "      <td>0.882924</td>\n",
              "      <td>0.021572</td>\n",
              "      <td>0.001236</td>\n",
              "      <td>0.002972</td>\n",
              "      <td>0.003366</td>\n",
              "      <td>True</td>\n",
              "      <td>0.877898</td>\n",
              "      <td>0.866465</td>\n",
              "      <td>0.001090</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.027492</td>\n",
              "      <td>0.043644</td>\n",
              "      <td>0.042874</td>\n",
              "      <td>0.027784</td>\n",
              "      <td>True</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.001236</td>\n",
              "      <td>0.001236</td>\n",
              "      <td>0.001236</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5349 rows × 21 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "               Close  marketRisk   returns    Target    Change  Signal  \\\n",
              "Date                                                                     \n",
              "1998-05-21  0.894694    0.017921 -0.005816 -0.001439 -0.001608   False   \n",
              "1998-05-22  0.893256    0.054887 -0.001608  0.004813  0.005388    True   \n",
              "1998-05-25  0.898069    0.072076  0.005388  0.001373  0.001529    True   \n",
              "1998-05-26  0.899442    0.120229  0.001529  0.006027  0.006700    True   \n",
              "1998-05-27  0.905469    0.057863  0.006700 -0.001718 -0.001898   False   \n",
              "1998-05-28  0.903751    0.039126 -0.001898  0.003361  0.003719    True   \n",
              "1998-05-29  0.907112    0.044906  0.003719 -0.004096 -0.004515   False   \n",
              "1998-06-01  0.903016    0.082701 -0.004515  0.000734  0.000813    True   \n",
              "1998-06-02  0.903751    0.031756  0.000813 -0.005278 -0.005840   False   \n",
              "1998-06-03  0.898473    0.005857 -0.005840  0.000000  0.000000   False   \n",
              "1998-06-04  0.898473    0.020549  0.000000  0.002428  0.002703    True   \n",
              "1998-06-05  0.900901    0.033363  0.002703  0.002441  0.002710    True   \n",
              "1998-06-08  0.903342    0.055741  0.002710 -0.000245 -0.000271   False   \n",
              "1998-06-09  0.903098    0.015388 -0.000271  0.008064  0.008929    True   \n",
              "1998-06-10  0.911162    0.049067  0.008929  0.006775  0.007435    True   \n",
              "1998-06-11  0.917936    0.067583  0.007435 -0.003191 -0.003476   False   \n",
              "1998-06-12  0.914746    0.057471 -0.003476  0.002013  0.002200    True   \n",
              "1998-06-15  0.916758    0.058427  0.002200 -0.006012 -0.006557   False   \n",
              "1998-06-16  0.910747    0.017422 -0.006557 -0.006506 -0.007144   False   \n",
              "1998-06-17  0.904241    0.018393 -0.007144  0.002460  0.002720    True   \n",
              "1998-06-18  0.906701    0.054887  0.002720 -0.002296 -0.002532   False   \n",
              "1998-06-19  0.904404    0.027351 -0.002532  0.003037  0.003358    True   \n",
              "1998-06-22  0.907441    0.061778  0.003358  0.003389  0.003734    True   \n",
              "1998-06-23  0.910830    0.055145  0.003734 -0.000332 -0.000364   False   \n",
              "1998-06-24  0.910498    0.049160 -0.000364 -0.002150 -0.002362   False   \n",
              "1998-06-25  0.908348    0.043022 -0.002362  0.005980  0.006583    True   \n",
              "1998-06-26  0.914328    0.052612  0.006583  0.000000  0.000000   False   \n",
              "1998-06-29  0.914328    0.021773  0.000000 -0.000835 -0.000913   False   \n",
              "1998-06-30  0.913492    0.021075 -0.000913  0.004781  0.005234    True   \n",
              "1998-07-01  0.918274    0.013918  0.005234  0.000928  0.001011    True   \n",
              "...              ...         ...       ...       ...       ...     ...   \n",
              "2018-10-16  0.864080    0.046064  0.000346  0.005561  0.006435    True   \n",
              "2018-10-17  0.869641    0.063355  0.006435  0.003569  0.004104    True   \n",
              "2018-10-18  0.873210    0.063601  0.004104 -0.004627 -0.005298   False   \n",
              "2018-10-19  0.868583    0.035059 -0.005298  0.003789  0.004362    True   \n",
              "2018-10-22  0.872372    0.063041  0.004362 -0.000456 -0.000523   False   \n",
              "2018-10-23  0.871916    0.024420 -0.000523  0.005970  0.006848    True   \n",
              "2018-10-24  0.877886    0.039890  0.006848  0.001312  0.001495    True   \n",
              "2018-10-25  0.879198    0.045768  0.001495 -0.002082 -0.002368   False   \n",
              "2018-10-26  0.877116    0.046744 -0.002368  0.002314  0.002638    True   \n",
              "2018-10-29  0.879430    0.060689  0.002638  0.002093  0.002380    True   \n",
              "2018-10-30  0.881523    0.037025  0.002380  0.002650  0.003006    True   \n",
              "2018-10-31  0.884173    0.065147  0.003006 -0.007519 -0.008504   False   \n",
              "2018-11-01  0.876655    0.038183 -0.008504  0.001694  0.001932    True   \n",
              "2018-11-02  0.878349    0.064924  0.001932 -0.001617 -0.001841   False   \n",
              "2018-11-05  0.876732    0.045471 -0.001841 -0.001535 -0.001750   False   \n",
              "2018-11-06  0.875197    0.026117 -0.001750  0.000153  0.000175    True   \n",
              "2018-11-07  0.875350    0.025385  0.000175  0.004777  0.005457    True   \n",
              "2018-11-08  0.880127    0.050508  0.005457  0.002174  0.002470    True   \n",
              "2018-11-09  0.882301    0.060366  0.002470  0.009203  0.010431    True   \n",
              "2018-11-12  0.891504    0.069956  0.010431 -0.005686 -0.006378   False   \n",
              "2018-11-13  0.885818    0.045668 -0.006378 -0.001488 -0.001680   False   \n",
              "2018-11-14  0.884330    0.056054 -0.001680 -0.001405 -0.001589   False   \n",
              "2018-11-15  0.882924    0.041407 -0.001589 -0.007114 -0.008057   False   \n",
              "2018-11-16  0.875810    0.029247 -0.008057 -0.002524 -0.002882   False   \n",
              "2018-11-19  0.873286    0.006334 -0.002882  0.006221  0.007124    True   \n",
              "2018-11-20  0.879507    0.033459  0.007124 -0.001004 -0.001142   False   \n",
              "2018-11-21  0.878503    0.030244 -0.001142 -0.001848 -0.002104   False   \n",
              "2018-11-22  0.876655    0.041783 -0.002104  0.005180  0.005908    True   \n",
              "2018-11-23  0.881834    0.031853  0.005908  0.001090  0.001236    True   \n",
              "2018-11-26  0.882924    0.021572  0.001236  0.002972  0.003366    True   \n",
              "\n",
              "             Close30  Close100        r1        r2       ...               r4  \\\n",
              "Date                                                     ...                    \n",
              "1998-05-21  0.905687  0.916241 -0.005234 -0.007833       ...        -0.031832   \n",
              "1998-05-22  0.904816  0.916045 -0.001439 -0.011721       ...        -0.034474   \n",
              "1998-05-25  0.904089  0.915792  0.004813 -0.005518       ...        -0.030695   \n",
              "1998-05-26  0.903746  0.915530  0.001373 -0.007587       ...        -0.030358   \n",
              "1998-05-27  0.903617  0.915352  0.006027 -0.001643       ...        -0.016785   \n",
              "1998-05-28  0.903403  0.915163 -0.001718 -0.001882       ...        -0.011916   \n",
              "1998-05-29  0.903251  0.915000  0.003361  0.007184       ...         0.004422   \n",
              "1998-06-01  0.903043  0.914801 -0.004096  0.008322       ...        -0.007233   \n",
              "1998-06-02  0.902986  0.914626  0.000734  0.010495       ...        -0.003938   \n",
              "1998-06-03  0.902681  0.914384 -0.005278  0.000403       ...        -0.021321   \n",
              "1998-06-04  0.902308  0.914103  0.000000 -0.000970       ...        -0.027968   \n",
              "1998-06-05  0.902107  0.913835  0.002428 -0.004568       ...        -0.027173   \n",
              "1998-06-08  0.902044  0.913581  0.002441 -0.000408       ...        -0.017892   \n",
              "1998-06-09  0.901880  0.913314 -0.000245 -0.004014       ...        -0.015345   \n",
              "1998-06-10  0.901968  0.913203  0.008064  0.008146       ...        -0.002080   \n",
              "1998-06-11  0.902304  0.913226  0.006775  0.014186       ...         0.011072   \n",
              "1998-06-12  0.902717  0.913346 -0.003191  0.016273       ...        -0.002854   \n",
              "1998-06-15  0.903248  0.913411  0.002013  0.018286       ...        -0.005666   \n",
              "1998-06-16  0.903711  0.913442 -0.006012  0.009846       ...        -0.006853   \n",
              "1998-06-17  0.903987  0.913286 -0.006506  0.000899       ...        -0.017418   \n",
              "1998-06-18  0.904282  0.913089  0.002460  0.003603       ...        -0.009470   \n",
              "1998-06-19  0.904448  0.912852 -0.002296 -0.006757       ...        -0.016236   \n",
              "1998-06-22  0.904593  0.912714  0.003037 -0.010495       ...        -0.011930   \n",
              "1998-06-23  0.904869  0.912638  0.003389 -0.003916       ...        -0.012787   \n",
              "1998-06-24  0.905053  0.912611 -0.000332 -0.006260       ...        -0.010991   \n",
              "1998-06-25  0.905212  0.912626 -0.002150 -0.002399       ...        -0.010517   \n",
              "1998-06-26  0.905455  0.912593  0.005980  0.010087       ...        -0.006483   \n",
              "1998-06-29  0.905696  0.912512  0.000000  0.007627       ...         0.005154   \n",
              "1998-06-30  0.905958  0.912471 -0.000835  0.009088       ...         0.003740   \n",
              "1998-07-01  0.906569  0.912437  0.004781  0.010833       ...        -0.001098   \n",
              "...              ...       ...       ...       ...       ...              ...   \n",
              "2018-10-16  0.861075  0.860146  0.000299 -0.003749       ...         0.012509   \n",
              "2018-10-17  0.861399  0.860267  0.005561 -0.000681       ...         0.021392   \n",
              "2018-10-18  0.861822  0.860445  0.003569  0.002812       ...         0.008682   \n",
              "2018-10-19  0.861918  0.860554 -0.004627  0.000377       ...         0.007034   \n",
              "2018-10-22  0.862241  0.860729  0.003789  0.009782       ...         0.011935   \n",
              "2018-10-23  0.862582  0.860914 -0.000456  0.007013       ...         0.008954   \n",
              "2018-10-24  0.863169  0.861199  0.005970  0.014104       ...         0.013582   \n",
              "2018-10-25  0.863956  0.861514  0.001312  0.015118       ...         0.017204   \n",
              "2018-10-26  0.864527  0.861786 -0.002082  0.007475       ...         0.019115   \n",
              "2018-10-29  0.865309  0.862093  0.002314  0.006220       ...         0.024875   \n",
              "2018-10-30  0.866118  0.862392  0.002093  0.012940       ...         0.022786   \n",
              "2018-10-31  0.867032  0.862752  0.002650  0.011801       ...         0.018522   \n",
              "2018-11-01  0.867945  0.862873 -0.007519  0.004739       ...         0.012126   \n",
              "2018-11-02  0.868852  0.863041  0.001694  0.000463       ...         0.022404   \n",
              "2018-11-05  0.869698  0.863204 -0.001617 -0.002467       ...         0.017551   \n",
              "2018-11-06  0.870553  0.863326 -0.001535 -0.001919       ...         0.017196   \n",
              "2018-11-07  0.871334  0.863437  0.000153 -0.004080       ...         0.017276   \n",
              "2018-11-08  0.872032  0.863618  0.004777 -0.001397       ...         0.024622   \n",
              "2018-11-09  0.872726  0.863861  0.002174 -0.001872       ...         0.030875   \n",
              "2018-11-12  0.873648  0.864230  0.009203  0.014849       ...         0.040368   \n",
              "2018-11-13  0.874305  0.864501 -0.005686  0.007469       ...         0.034174   \n",
              "2018-11-14  0.874737  0.864688 -0.001488  0.007598       ...         0.027578   \n",
              "2018-11-15  0.875215  0.864872 -0.001405  0.007727       ...         0.026100   \n",
              "2018-11-16  0.875481  0.865071 -0.007114  0.000460       ...         0.020012   \n",
              "2018-11-19  0.875580  0.865212 -0.002524 -0.006841       ...         0.019242   \n",
              "2018-11-20  0.875883  0.865427  0.006221 -0.002794       ...         0.021801   \n",
              "2018-11-21  0.876226  0.865631 -0.001004 -0.013001       ...         0.019249   \n",
              "2018-11-22  0.876695  0.865843 -0.001848 -0.009163       ...         0.017622   \n",
              "2018-11-23  0.877260  0.866147  0.005180 -0.002495       ...         0.028446   \n",
              "2018-11-26  0.877898  0.866465  0.001090  0.000000       ...         0.027492   \n",
              "\n",
              "            marketrisk_avg30  marketrisk_avg90  marketrisk_avg7  \\\n",
              "Date                                                              \n",
              "1998-05-21          0.030184          0.034995         0.031504   \n",
              "1998-05-22          0.030500          0.035283         0.035738   \n",
              "1998-05-25          0.031202          0.035543         0.042474   \n",
              "1998-05-26          0.034618          0.036311         0.056034   \n",
              "1998-05-27          0.036330          0.036533         0.055463   \n",
              "1998-05-28          0.035828          0.036895         0.055423   \n",
              "1998-05-29          0.036237          0.037338         0.058144   \n",
              "1998-06-01          0.038099          0.037892         0.067398   \n",
              "1998-06-02          0.038561          0.038163         0.064094   \n",
              "1998-06-03          0.038295          0.037822         0.054634   \n",
              "1998-06-04          0.037835          0.037518         0.040394   \n",
              "1998-06-05          0.038003          0.037519         0.036894   \n",
              "1998-06-08          0.037301          0.037236         0.039268   \n",
              "1998-06-09          0.036329          0.036582         0.035051   \n",
              "1998-06-10          0.036190          0.036510         0.030246   \n",
              "1998-06-11          0.037385          0.036613         0.035364   \n",
              "1998-06-12          0.038482          0.037026         0.042737   \n",
              "1998-06-15          0.039980          0.037746         0.048149   \n",
              "1998-06-16          0.040951          0.037204         0.045871   \n",
              "1998-06-17          0.040985          0.037192         0.040536   \n",
              "1998-06-18          0.042129          0.037509         0.046179   \n",
              "1998-06-19          0.041939          0.037346         0.043076   \n",
              "1998-06-22          0.043055          0.037638         0.042247   \n",
              "1998-06-23          0.044083          0.037913         0.041915   \n",
              "1998-06-24          0.044880          0.038555         0.040591   \n",
              "1998-06-25          0.045483          0.038685         0.044248   \n",
              "1998-06-26          0.046393          0.038958         0.049136   \n",
              "1998-06-29          0.045057          0.038731         0.044406   \n",
              "1998-06-30          0.044446          0.038792         0.043509   \n",
              "1998-07-01          0.044048          0.038531         0.036672   \n",
              "...                      ...               ...              ...   \n",
              "2018-10-16          0.047195          0.043938         0.043278   \n",
              "2018-10-17          0.047701          0.044340         0.044644   \n",
              "2018-10-18          0.049071          0.044886         0.048360   \n",
              "2018-10-19          0.048989          0.044557         0.046086   \n",
              "2018-10-22          0.049574          0.044681         0.051670   \n",
              "2018-10-23          0.048726          0.044506         0.045925   \n",
              "2018-10-24          0.049034          0.044314         0.047919   \n",
              "2018-10-25          0.049487          0.044216         0.047876   \n",
              "2018-10-26          0.049385          0.044216         0.045503   \n",
              "2018-10-29          0.050389          0.044624         0.045087   \n",
              "2018-10-30          0.050033          0.044629         0.045368   \n",
              "2018-10-31          0.050838          0.044868         0.045669   \n",
              "2018-11-01          0.051000          0.044865         0.047635   \n",
              "2018-11-02          0.051805          0.044856         0.051211   \n",
              "2018-11-05          0.051697          0.044835         0.051169   \n",
              "2018-11-06          0.050886          0.044647         0.048222   \n",
              "2018-11-07          0.049994          0.044518         0.043179   \n",
              "2018-11-08          0.049714          0.044651         0.045105   \n",
              "2018-11-09          0.049158          0.044969         0.044422   \n",
              "2018-11-12          0.049569          0.045400         0.048961   \n",
              "2018-11-13          0.049300          0.045223         0.046210   \n",
              "2018-11-14          0.048751          0.045235         0.047722   \n",
              "2018-11-15          0.047424          0.044973         0.049906   \n",
              "2018-11-16          0.046699          0.044435         0.050458   \n",
              "2018-11-19          0.045117          0.043940         0.044147   \n",
              "2018-11-20          0.044979          0.043843         0.040304   \n",
              "2018-11-21          0.044288          0.043271         0.034630   \n",
              "2018-11-22          0.044882          0.043111         0.034075   \n",
              "2018-11-23          0.043790          0.043037         0.030618   \n",
              "2018-11-26          0.043644          0.042874         0.027784   \n",
              "\n",
              "            predictSignal predictedweight  labelledweight         r  \\\n",
              "Date                                                                  \n",
              "1998-05-21              0              -1              -1       NaN   \n",
              "1998-05-22              0              -1               1 -0.001608   \n",
              "1998-05-25              0              -1               1  0.005388   \n",
              "1998-05-26              0              -1               1  0.001529   \n",
              "1998-05-27              0              -1              -1  0.006700   \n",
              "1998-05-28              0              -1               1 -0.001898   \n",
              "1998-05-29              0              -1              -1  0.003719   \n",
              "1998-06-01              0              -1               1 -0.004515   \n",
              "1998-06-02              0              -1              -1  0.000813   \n",
              "1998-06-03              0              -1              -1 -0.005840   \n",
              "1998-06-04              0              -1               1  0.000000   \n",
              "1998-06-05              0              -1               1  0.002703   \n",
              "1998-06-08              0              -1              -1  0.002710   \n",
              "1998-06-09              0              -1               1 -0.000271   \n",
              "1998-06-10              0              -1               1  0.008929   \n",
              "1998-06-11              0              -1              -1  0.007435   \n",
              "1998-06-12              0              -1               1 -0.003476   \n",
              "1998-06-15              0              -1              -1  0.002200   \n",
              "1998-06-16              0              -1              -1 -0.006557   \n",
              "1998-06-17              0              -1               1 -0.007144   \n",
              "1998-06-18              0              -1              -1  0.002720   \n",
              "1998-06-19              0              -1               1 -0.002532   \n",
              "1998-06-22              0              -1               1  0.003358   \n",
              "1998-06-23              0              -1              -1  0.003734   \n",
              "1998-06-24              0              -1              -1 -0.000364   \n",
              "1998-06-25              0              -1               1 -0.002362   \n",
              "1998-06-26              0              -1              -1  0.006583   \n",
              "1998-06-29              0              -1              -1  0.000000   \n",
              "1998-06-30              0              -1               1 -0.000913   \n",
              "1998-07-01              0              -1               1  0.005234   \n",
              "...                   ...             ...             ...       ...   \n",
              "2018-10-16           True               1               1  0.000346   \n",
              "2018-10-17           True               1               1  0.006435   \n",
              "2018-10-18           True               1              -1  0.004104   \n",
              "2018-10-19           True               1               1 -0.005298   \n",
              "2018-10-22           True               1              -1  0.004362   \n",
              "2018-10-23           True               1               1 -0.000523   \n",
              "2018-10-24           True               1               1  0.006848   \n",
              "2018-10-25           True               1              -1  0.001495   \n",
              "2018-10-26           True               1               1 -0.002368   \n",
              "2018-10-29           True               1               1  0.002638   \n",
              "2018-10-30           True               1               1  0.002380   \n",
              "2018-10-31           True               1              -1  0.003006   \n",
              "2018-11-01           True               1               1 -0.008504   \n",
              "2018-11-02           True               1              -1  0.001932   \n",
              "2018-11-05           True               1              -1 -0.001841   \n",
              "2018-11-06           True               1               1 -0.001750   \n",
              "2018-11-07           True               1               1  0.000175   \n",
              "2018-11-08           True               1               1  0.005457   \n",
              "2018-11-09           True               1               1  0.002470   \n",
              "2018-11-12           True               1              -1  0.010431   \n",
              "2018-11-13           True               1              -1 -0.006378   \n",
              "2018-11-14           True               1              -1 -0.001680   \n",
              "2018-11-15           True               1              -1 -0.001589   \n",
              "2018-11-16           True               1              -1 -0.008057   \n",
              "2018-11-19           True               1               1 -0.002882   \n",
              "2018-11-20           True               1              -1  0.007124   \n",
              "2018-11-21           True               1              -1 -0.001142   \n",
              "2018-11-22           True               1               1 -0.002104   \n",
              "2018-11-23           True               1               1  0.005908   \n",
              "2018-11-26           True               1               1  0.001236   \n",
              "\n",
              "            predictedreturns  labelledreturns  \n",
              "Date                                           \n",
              "1998-05-21               NaN              NaN  \n",
              "1998-05-22          0.001608         0.001608  \n",
              "1998-05-25         -0.005388         0.005388  \n",
              "1998-05-26         -0.001529         0.001529  \n",
              "1998-05-27         -0.006700         0.006700  \n",
              "1998-05-28          0.001898         0.001898  \n",
              "1998-05-29         -0.003719         0.003719  \n",
              "1998-06-01          0.004515         0.004515  \n",
              "1998-06-02         -0.000813         0.000813  \n",
              "1998-06-03          0.005840         0.005840  \n",
              "1998-06-04         -0.000000        -0.000000  \n",
              "1998-06-05         -0.002703         0.002703  \n",
              "1998-06-08         -0.002710         0.002710  \n",
              "1998-06-09          0.000271         0.000271  \n",
              "1998-06-10         -0.008929         0.008929  \n",
              "1998-06-11         -0.007435         0.007435  \n",
              "1998-06-12          0.003476         0.003476  \n",
              "1998-06-15         -0.002200         0.002200  \n",
              "1998-06-16          0.006557         0.006557  \n",
              "1998-06-17          0.007144         0.007144  \n",
              "1998-06-18         -0.002720         0.002720  \n",
              "1998-06-19          0.002532         0.002532  \n",
              "1998-06-22         -0.003358         0.003358  \n",
              "1998-06-23         -0.003734         0.003734  \n",
              "1998-06-24          0.000364         0.000364  \n",
              "1998-06-25          0.002362         0.002362  \n",
              "1998-06-26         -0.006583         0.006583  \n",
              "1998-06-29         -0.000000        -0.000000  \n",
              "1998-06-30          0.000913         0.000913  \n",
              "1998-07-01         -0.005234         0.005234  \n",
              "...                      ...              ...  \n",
              "2018-10-16          0.000346         0.000346  \n",
              "2018-10-17          0.006435         0.006435  \n",
              "2018-10-18          0.004104         0.004104  \n",
              "2018-10-19         -0.005298         0.005298  \n",
              "2018-10-22          0.004362         0.004362  \n",
              "2018-10-23         -0.000523         0.000523  \n",
              "2018-10-24          0.006848         0.006848  \n",
              "2018-10-25          0.001495         0.001495  \n",
              "2018-10-26         -0.002368         0.002368  \n",
              "2018-10-29          0.002638         0.002638  \n",
              "2018-10-30          0.002380         0.002380  \n",
              "2018-10-31          0.003006         0.003006  \n",
              "2018-11-01         -0.008504         0.008504  \n",
              "2018-11-02          0.001932         0.001932  \n",
              "2018-11-05         -0.001841         0.001841  \n",
              "2018-11-06         -0.001750         0.001750  \n",
              "2018-11-07          0.000175         0.000175  \n",
              "2018-11-08          0.005457         0.005457  \n",
              "2018-11-09          0.002470         0.002470  \n",
              "2018-11-12          0.010431         0.010431  \n",
              "2018-11-13         -0.006378         0.006378  \n",
              "2018-11-14         -0.001680         0.001680  \n",
              "2018-11-15         -0.001589         0.001589  \n",
              "2018-11-16         -0.008057         0.008057  \n",
              "2018-11-19         -0.002882         0.002882  \n",
              "2018-11-20          0.007124         0.007124  \n",
              "2018-11-21         -0.001142         0.001142  \n",
              "2018-11-22         -0.002104         0.002104  \n",
              "2018-11-23          0.005908         0.005908  \n",
              "2018-11-26          0.001236         0.001236  \n",
              "\n",
              "[5349 rows x 21 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 96
        }
      ]
    },
    {
      "metadata": {
        "id": "XP6EtlYYSjdd",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# define confusion matrix function\n",
        "def plot_confusion_matrix(cm, classes,\n",
        "                          normalize=False,\n",
        "                          title='Confusion matrix',\n",
        "                          cmap=plt.cm.Blues):\n",
        "    \"\"\"\n",
        "    This function prints and plots the confusion matrix.\n",
        "    Normalization can be applied by setting `normalize=True`.\n",
        "    \"\"\"\n",
        "    if normalize:\n",
        "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
        "        print(\"Normalized confusion matrix\")\n",
        "    else:\n",
        "        print('Confusion matrix, without normalization')\n",
        "\n",
        "    print(cm)\n",
        "\n",
        "#     plt.figure(figsize= (10,10))\n",
        "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
        "    plt.title(title)\n",
        "#     plt.colorbar()\n",
        "    tick_marks = np.arange(len(classes))\n",
        "    plt.xticks(tick_marks, classes, rotation=45)\n",
        "    plt.yticks(tick_marks, classes)\n",
        "\n",
        "    fmt = '.2f' if normalize else 'd'\n",
        "    thresh = cm.max() / 2.\n",
        "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
        "        plt.text(j, i, format(cm[i, j], fmt),\n",
        "                 horizontalalignment=\"center\",\n",
        "                 color=\"black\" if cm[i, j] > thresh else \"black\")\n",
        "\n",
        "    plt.ylabel('True label')\n",
        "    plt.xlabel('Predicted label')\n",
        "    plt.tight_layout()\n",
        "    \n",
        "def plotConfusion(start, end, string):\n",
        "  real = tradeTable[\"labelledweight\"]\n",
        "  pred = tradeTable[\"predictedweight\"]\n",
        "  cnf_matrix = confusion_matrix(real[start:end], pred[start:end])\n",
        "  accuracy = np.around(accuracy_score(real[start:end], pred[start:end]),2)\n",
        "  f1= np.around(f1_score(real[start:end], pred[start:end], average='binary'), 2)\n",
        "  auc = np.around(areauc(real[start:end], pred[start:end]), 2)\n",
        "  prec = np.around(precision_score(real[start:end], pred[start:end], average='binary'), 2)\n",
        "  recall = np.around(recall_score(real[start:end], pred[start:end], average='binary'), 2)\n",
        "  mcc =  np.around(matthews_corrcoef(real[start:end], pred[start:end]), 2)\n",
        "  # np.set_printoptions(precision=2)\n",
        "\n",
        "\n",
        "  plot_confusion_matrix(cnf_matrix, classes=['Sell', 'buy'],\n",
        "                        title=string + ' Accuracy: ' +str(accuracy)\n",
        "                        + ', AUC:' + str(auc)\n",
        "#                         + ', F1 Score:' + str(f1)\n",
        "#                         + ', Precison:' + str(prec)\n",
        "#                         + ', Recall:' + str(recall)\n",
        "                        + ', MCC:' + str(mcc))"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}