Iris_Dataset.ipynb

{
  "cells": [
    {
      "metadata": {
        "_uuid": "e0db02d257c6bb6631773d9bcd03ff022e25649a"
      },
      "cell_type": "markdown",
      "source": "<img src=\"http://s5047.pcdn.co/wp-content/uploads/2015/04/iris_petal_sepal.png\">\n"
    },
    {
      "metadata": {
        "_uuid": "3c23607dc3ccb91848089096ccfd7620666903ae"
      },
      "cell_type": "markdown",
      "source": "The Iris dataset was used in R.A. Fisher's classic 1936 paper, The Use of Multiple Measurements in Taxonomic Problems, and can also be found on the UCI Machine Learning Repository.\n\nIt includes three iris species with 50 samples each as well as some properties about each flower. One flower species is linearly separable from the other two, but the other two are not linearly separable from each other.\n\nThe link for dataset - https://www.kaggle.com/uciml/iris"
    },
    {
      "metadata": {
        "_kg_hide-input": true,
        "trusted": true,
        "_uuid": "ab3c2f7c07531992dd024b7d1df6997ba1a251b3"
      },
      "cell_type": "code",
      "source": "import numpy as np\nimport random\nimport pandas as pd\nfrom pandas.tools import plotting\nimport seaborn as sns\nimport matplotlib.pyplot as plt\n%matplotlib inline\n\nimport plotly.offline as py\nfrom plotly.offline import init_notebook_mode, iplot\nimport plotly.graph_objs as go\nfrom plotly import tools\ninit_notebook_mode(connected=True)  \nimport plotly.figure_factory as ff\n\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.svm import SVC\nfrom sklearn.naive_bayes import GaussianNB\nfrom sklearn.tree import DecisionTreeClassifier\nfrom sklearn.ensemble import RandomForestClassifier, ExtraTreesClassifier\nfrom sklearn.neighbors import KNeighborsClassifier\n# from lightgbm import LGBMClassifier\nfrom sklearn.metrics import  accuracy_score\n\nimport xgboost as xgb\nfrom catboost import CatBoostClassifier\n\nfrom sklearn.preprocessing import StandardScaler, LabelBinarizer\n# auxiliary function\nfrom sklearn.preprocessing import LabelEncoder\ndef random_colors(number_of_colors):\n    color = [\"#\"+''.join([random.choice('0123456789ABCDEF') for j in range(6)])\n                 for i in range(number_of_colors)]\n    return color\n\nimport warnings\nwarnings.filterwarnings('ignore')",
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": "<script>requirejs.config({paths: { 'plotly': ['https://cdn.plot.ly/plotly-latest.min']},});if(!window.Plotly) {{require(['plotly'],function(plotly) {window.Plotly=plotly;});}}</script>",
            "text/vnd.plotly.v1+html": "<script>requirejs.config({paths: { 'plotly': ['https://cdn.plot.ly/plotly-latest.min']},});if(!window.Plotly) {{require(['plotly'],function(plotly) {window.Plotly=plotly;});}}</script>"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "cc720b9cb2ab8ee436074089b4d6d74fc5b8f882"
      },
      "cell_type": "markdown",
      "source": "Load the dataset in memory using pandas library."
    },
    {
      "metadata": {
        "_kg_hide-input": true,
        "trusted": true,
        "_uuid": "0caff7f827cd7a8a0352992e2a80d42fdb8d26dc"
      },
      "cell_type": "code",
      "source": "df = pd.read_csv('../input/Iris.csv')\ntable = ff.create_table(df.head())",
      "execution_count": 2,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "bc136f6f485d14579bcb3bb41c691b288d534b41"
      },
      "cell_type": "code",
      "source": "df.head()",
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 3,
          "data": {
            "text/plain": "   Id  SepalLengthCm     ...       PetalWidthCm      Species\n0   1            5.1     ...                0.2  Iris-setosa\n1   2            4.9     ...                0.2  Iris-setosa\n2   3            4.7     ...                0.2  Iris-setosa\n3   4            4.6     ...                0.2  Iris-setosa\n4   5            5.0     ...                0.2  Iris-setosa\n\n[5 rows x 6 columns]",
            "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>Id</th>\n      <th>SepalLengthCm</th>\n      <th>SepalWidthCm</th>\n      <th>PetalLengthCm</th>\n      <th>PetalWidthCm</th>\n      <th>Species</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>1</td>\n      <td>5.1</td>\n      <td>3.5</td>\n      <td>1.4</td>\n      <td>0.2</td>\n      <td>Iris-setosa</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>2</td>\n      <td>4.9</td>\n      <td>3.0</td>\n      <td>1.4</td>\n      <td>0.2</td>\n      <td>Iris-setosa</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>3</td>\n      <td>4.7</td>\n      <td>3.2</td>\n      <td>1.3</td>\n      <td>0.2</td>\n      <td>Iris-setosa</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>4</td>\n      <td>4.6</td>\n      <td>3.1</td>\n      <td>1.5</td>\n      <td>0.2</td>\n      <td>Iris-setosa</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>5</td>\n      <td>5.0</td>\n      <td>3.6</td>\n      <td>1.4</td>\n      <td>0.2</td>\n      <td>Iris-setosa</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "_kg_hide-input": true,
        "trusted": true,
        "_uuid": "1179c41df2d7a6e61012afbeab13b4cb97003724"
      },
      "cell_type": "code",
      "source": "df.info()",
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 150 entries, 0 to 149\nData columns (total 6 columns):\nId               150 non-null int64\nSepalLengthCm    150 non-null float64\nSepalWidthCm     150 non-null float64\nPetalLengthCm    150 non-null float64\nPetalWidthCm     150 non-null float64\nSpecies          150 non-null object\ndtypes: float64(4), int64(1), object(1)\nmemory usage: 7.1+ KB\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "d94cacaeb20745de52b44cecadb9f124bba6ede0"
      },
      "cell_type": "markdown",
      "source": "so there is no null values available in the data set."
    },
    {
      "metadata": {
        "_uuid": "00fb91122216780bf8fbd66c2a13cee1eddfade0"
      },
      "cell_type": "markdown",
      "source": " **Visualization of the Dataset**"
    },
    {
      "metadata": {
        "_uuid": "3771de617effaadbf899a4bcbf7db24ec45e9402"
      },
      "cell_type": "markdown",
      "source": "given the coloums are<br>\nSepalLengthCm<br>\nSepalWidthCm<br>\nPetalLengthCm<br>\nPetalWidthCm<br>\nSpecies<br>"
    },
    {
      "metadata": {
        "_uuid": "4c54df7ef24e7d9e75fb6f2e6b137ef086a06319"
      },
      "cell_type": "markdown",
      "source": "**Types of Species**"
    },
    {
      "metadata": {
        "_kg_hide-input": true,
        "trusted": true,
        "_uuid": "ceea7a2606503e6afa020ec014be2efd989fd9d4"
      },
      "cell_type": "code",
      "source": "Species = df['Species'].unique()\nSpecies",
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 5,
          "data": {
            "text/plain": "array(['Iris-setosa', 'Iris-versicolor', 'Iris-virginica'], dtype=object)"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "3deeea05193d2bff16945fa5eee75f57144ecf06"
      },
      "cell_type": "markdown",
      "source": "So there are three types of species \nIris-setosa<br/>\nIris-versicolor<br/>\nIris-virginica<br/>"
    },
    {
      "metadata": {
        "_uuid": "7e7c935c6749f76b8bb619b1a629861660d5878d"
      },
      "cell_type": "markdown",
      "source": "Corelation between features, They seem pretty independent."
    },
    {
      "metadata": {
        "_kg_hide-input": true,
        "trusted": true,
        "_uuid": "47326e3e9bc656c6dbf3abffaddd8464180ba585"
      },
      "cell_type": "code",
      "source": "corelation = df.corr()\ndata = [go.Heatmap(z = np.array(corelation.values),\n                   x = np.array(corelation.columns),\n                   y = np.array(corelation.columns),\n                     colorscale='Blackbody',)\n          ]\npy.iplot(data)",
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.plotly.v1+json": {
              "data": [
                {
                  "colorscale": "Blackbody",
                  "x": [
                    "Id",
                    "SepalLengthCm",
                    "SepalWidthCm",
                    "PetalLengthCm",
                    "PetalWidthCm"
                  ],
                  "y": [
                    "Id",
                    "SepalLengthCm",
                    "SepalWidthCm",
                    "PetalLengthCm",
                    "PetalWidthCm"
                  ],
                  "z": [
                    [
                      1,
                      0.7166762728539001,
                      -0.397728811465928,
                      0.8827473181390559,
                      0.8997585770933502
                    ],
                    [
                      0.7166762728539001,
                      1,
                      -0.10936924995064935,
                      0.8717541573048719,
                      0.8179536333691635
                    ],
                    [
                      -0.397728811465928,
                      -0.10936924995064935,
                      1,
                      -0.42051609640115484,
                      -0.3565440896138055
                    ],
                    [
                      0.8827473181390559,
                      0.8717541573048719,
                      -0.42051609640115484,
                      1,
                      0.9627570970509667
                    ],
                    [
                      0.8997585770933502,
                      0.8179536333691635,
                      -0.3565440896138055,
                      0.9627570970509667,
                      1
                    ]
                  ],
                  "type": "heatmap",
                  "uid": "6f8241d2-c2fc-11e8-94b0-0242ac130002"
                }
              ],
              "layout": {}
            },
            "text/html": "<div id=\"85e49837-a9c0-41a6-84ed-77c53dffd250\" style=\"height: 525px; width: 100%;\" class=\"plotly-graph-div\"></div><script type=\"text/javascript\">require([\"plotly\"], function(Plotly) { window.PLOTLYENV=window.PLOTLYENV || {};window.PLOTLYENV.BASE_URL=\"https://plot.ly\";Plotly.newPlot(\"85e49837-a9c0-41a6-84ed-77c53dffd250\", [{\"colorscale\": \"Blackbody\", \"x\": [\"Id\", \"SepalLengthCm\", \"SepalWidthCm\", \"PetalLengthCm\", \"PetalWidthCm\"], \"y\": [\"Id\", \"SepalLengthCm\", \"SepalWidthCm\", \"PetalLengthCm\", \"PetalWidthCm\"], \"z\": [[1.0, 0.7166762728539001, -0.397728811465928, 0.8827473181390559, 0.8997585770933502], [0.7166762728539001, 1.0, -0.10936924995064935, 0.8717541573048719, 0.8179536333691635], [-0.397728811465928, -0.10936924995064935, 1.0, -0.42051609640115484, -0.3565440896138055], [0.8827473181390559, 0.8717541573048719, -0.42051609640115484, 1.0, 0.9627570970509667], [0.8997585770933502, 0.8179536333691635, -0.3565440896138055, 0.9627570970509667, 1.0]], \"type\": \"heatmap\", \"uid\": \"6f84c89e-c2fc-11e8-94b0-0242ac130002\"}], {}, {\"showLink\": true, \"linkText\": \"Export to plot.ly\"})});</script>",
            "text/vnd.plotly.v1+html": "<div id=\"85e49837-a9c0-41a6-84ed-77c53dffd250\" style=\"height: 525px; width: 100%;\" class=\"plotly-graph-div\"></div><script type=\"text/javascript\">require([\"plotly\"], function(Plotly) { window.PLOTLYENV=window.PLOTLYENV || {};window.PLOTLYENV.BASE_URL=\"https://plot.ly\";Plotly.newPlot(\"85e49837-a9c0-41a6-84ed-77c53dffd250\", [{\"colorscale\": \"Blackbody\", \"x\": [\"Id\", \"SepalLengthCm\", \"SepalWidthCm\", \"PetalLengthCm\", \"PetalWidthCm\"], \"y\": [\"Id\", \"SepalLengthCm\", \"SepalWidthCm\", \"PetalLengthCm\", \"PetalWidthCm\"], \"z\": [[1.0, 0.7166762728539001, -0.397728811465928, 0.8827473181390559, 0.8997585770933502], [0.7166762728539001, 1.0, -0.10936924995064935, 0.8717541573048719, 0.8179536333691635], [-0.397728811465928, -0.10936924995064935, 1.0, -0.42051609640115484, -0.3565440896138055], [0.8827473181390559, 0.8717541573048719, -0.42051609640115484, 1.0, 0.9627570970509667], [0.8997585770933502, 0.8179536333691635, -0.3565440896138055, 0.9627570970509667, 1.0]], \"type\": \"heatmap\", \"uid\": \"6f84c89e-c2fc-11e8-94b0-0242ac130002\"}], {}, {\"showLink\": true, \"linkText\": \"Export to plot.ly\"})});</script>"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "6dd6113a725c65cd882b9694c928ae50cf2108c8"
      },
      "cell_type": "markdown",
      "source": "Lets start with some traditional ML."
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "174f80946174bd6bc0ba64bfebf6a5c8fac36db8"
      },
      "cell_type": "code",
      "source": "x = df[['SepalLengthCm','SepalWidthCm','PetalLengthCm','PetalWidthCm']]\ny = df['Species']",
      "execution_count": 7,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "6150c9fdda1b523ba908271960aa4f6508bffe6d"
      },
      "cell_type": "code",
      "source": "encoder = LabelEncoder()\ny = encoder.fit_transform(y)",
      "execution_count": 8,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "8c130a4ad47d1336614303a58d21b3a498b203e2"
      },
      "cell_type": "code",
      "source": "y",
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 9,
          "data": {
            "text/plain": "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n       0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "b1133179fff6a668edc493033557b4ab0dbbc65a"
      },
      "cell_type": "markdown",
      "source": "As you can see Iris-setosa Iris-versicolor Iris-virginica are converted to 0, 1, 2 respectively.\nFirst we are splitting the data set into training data and testing data which is 7:3 ratio. "
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "e95d548a7f9a3fbef363cf50060375e3bd4b2def"
      },
      "cell_type": "code",
      "source": "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 0.2, random_state = 42)",
      "execution_count": 10,
      "outputs": []
    },
    {
      "metadata": {
        "_uuid": "a429c38170278cbc2b6220b9e7b2cdcf97104c75"
      },
      "cell_type": "markdown",
      "source": "**List of algorithms**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "a37b694d493d259be10a23e1123f2dd0a349e12f"
      },
      "cell_type": "markdown",
      "source": "Since it is a classification problem we will be using<br>\nLogistic regression<br>\nDecision tree<br>\nKNN<br>\nSVM<br>\nNaive Bayes Classification<br>\nRandom forest<br>\nGBM<br>\nLigthGBM<br>"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "b222fe6d3fc4d877e3d428d6a2ff8a5d29b2c528"
      },
      "cell_type": "markdown",
      "source": "**Logistic regression**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "ce27be1c4fe5918c150798b398d3004969b5ba0f"
      },
      "cell_type": "code",
      "source": "lr_model = LogisticRegression()\nlr_model.fit(x_train,y_train)\nlr_predict = lr_model.predict(x_test)\nprint('Logistic Regression - ',accuracy_score(lr_predict,y_test))",
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Logistic Regression -  1.0\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "67fc07be60779c31c2ada8c725289d8ef943c1bd"
      },
      "cell_type": "markdown",
      "source": "**SVM**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "9c86b283ce845304587a40a9e92e22130b43caee"
      },
      "cell_type": "code",
      "source": "svm_model = SVC(kernel='linear')\nsvm_model.fit(x_train,y_train)\nsvc_predict = svm_model.predict(x_test)\nprint('SVM - ',accuracy_score(lr_predict,y_test))",
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": "SVM -  1.0\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "4e7abe757465fb5904e59acf6d5c49f37ff0e740"
      },
      "cell_type": "markdown",
      "source": "**Naive Bayes Classifier**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "ad312dc9438ae6a0afde56012bf3e101064373fe"
      },
      "cell_type": "code",
      "source": "nb_model = GaussianNB()\nnb_model.fit(x_train,y_train)\nnb_predict = nb_model.predict(x_test)\nprint('Naive bayes - ',accuracy_score(nb_predict,y_test))",
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Naive bayes -  1.0\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "4dda5f1d31e579cf25e2c18175cae01704053e76"
      },
      "cell_type": "markdown",
      "source": "**Decision Tree**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "567f22ed14b3fb568c4a85af9b306214108de7f4"
      },
      "cell_type": "code",
      "source": "dt_model = DecisionTreeClassifier(max_leaf_nodes=3)\ndt_model.fit(x_train,y_train)\ndt_predict = dt_model.predict(x_test)\nprint('Decision Tree - ',accuracy_score(dt_predict,y_test))",
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Decision Tree -  0.9666666666666667\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "42d55936cf6cbaa54948a268c9e243c733ce39cd"
      },
      "cell_type": "markdown",
      "source": "**Random Forest Classifier**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "4aeb3ce119c947d44b85dd398afaeb21504a2221"
      },
      "cell_type": "code",
      "source": "rfc_model = RandomForestClassifier(max_depth=3)\nrfc_model.fit(x_train,y_train)\nrfc_predict = rfc_model.predict(x_test)\nprint('Random Forest - ',accuracy_score(rfc_predict,y_test))",
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Random Forest -  1.0\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "46dc2d7de36b297dad987cdc98bb9cf41c29efd7"
      },
      "cell_type": "markdown",
      "source": "**Extra Trees Classifier**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "fc1c65b5b7b1eabbb09db35cec298b29ae00ec51"
      },
      "cell_type": "code",
      "source": "etc_model = ExtraTreesClassifier()\netc_model.fit(x_train,y_train)\netc_predict = etc_model.predict(x_test)\nprint('Extra Tree Classifier - ',accuracy_score(etc_predict,y_test))",
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Extra Tree Classifier -  0.9666666666666667\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "f575634697a8321b9fbc53a69025b8d21baa3a1c"
      },
      "cell_type": "markdown",
      "source": "**KNN**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "94a5fe5f4c50210bbc141d75cba423f0dde1b3b2"
      },
      "cell_type": "code",
      "source": "knn_model = KNeighborsClassifier(n_neighbors=3)\nknn_model.fit(x_train,y_train)\nknn_predict = knn_model.predict(x_test)\nprint('knn - ',accuracy_score(knn_predict,y_test))",
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": "knn -  1.0\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "4705805636047c4ce43898c0c36259407f543a0d"
      },
      "cell_type": "markdown",
      "source": "**Deep Learning**"
    },
    {
      "metadata": {
        "_uuid": "68b1d94d10fe279961e28e6bba90f27baeceebe2"
      },
      "cell_type": "markdown",
      "source": "Best place to understand deep learning.\nPlease follow the blog\nhttps://machinelearningmastery.com/what-is-deep-learning/"
    },
    {
      "metadata": {
        "_kg_hide-input": true,
        "trusted": true,
        "_uuid": "2bc41a138e69a3d02acdfa46cd47ad2b4fcfa74e"
      },
      "cell_type": "code",
      "source": "import keras\nfrom keras.models import Sequential\nfrom keras.layers import Dense",
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Using TensorFlow backend.\n",
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "_kg_hide-input": true,
        "trusted": true,
        "_uuid": "93256144af33cc23c71dd8336f5622b2aab2ef90"
      },
      "cell_type": "code",
      "source": "from sklearn.preprocessing import StandardScaler, LabelBinarizer\nX = df[['SepalLengthCm', 'SepalWidthCm', 'PetalLengthCm', 'PetalWidthCm']]\ny = df['Species']\n\nX = StandardScaler().fit_transform(X)\ny = LabelBinarizer().fit_transform(y)",
      "execution_count": 19,
      "outputs": []
    },
    {
      "metadata": {
        "_uuid": "9df7b248a5352b8e2d0e1d9d3ed68d9a253c02e2"
      },
      "cell_type": "markdown",
      "source": "Spliting the data into train - 80% and test - 20%"
    },
    {
      "metadata": {
        "_kg_hide-input": true,
        "trusted": true,
        "_uuid": "8984cfef81f4e31db498376f5ca2bf72748faf13"
      },
      "cell_type": "code",
      "source": "x_train, x_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42)",
      "execution_count": 20,
      "outputs": []
    },
    {
      "metadata": {
        "_uuid": "64eecf88ba78f8616345496654ab70a04019f698"
      },
      "cell_type": "markdown",
      "source": "**Deep learning**"
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "77cded2a0a730f5a6c3717a10449c890e4195f73"
      },
      "cell_type": "code",
      "source": "deep_model = Sequential()\ndeep_model.add(Dense( 4, input_dim=4, activation = 'relu'))\ndeep_model.add(Dense( units = 100, activation= 'relu'))\ndeep_model.add(Dense( units = 100, activation= 'relu'))\ndeep_model.add(Dense( units = 100, activation= 'relu'))\ndeep_model.add(Dense( units = 100, activation= 'relu'))\ndeep_model.add(Dense( units = 100, activation= 'relu'))\ndeep_model.add(Dense( units = 100, activation= 'relu'))\ndeep_model.add(Dense( units = 100, activation= 'relu'))\ndeep_model.add(Dense( units = 100, activation= 'relu'))\ndeep_model.add(Dense( units = 3, activation= 'softmax'))\ndeep_model.compile(loss = 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])\n",
      "execution_count": 21,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "4be5a0688349325ac9f08365c28265571770dfe0"
      },
      "cell_type": "code",
      "source": "deep_history = deep_model.fit(x_train, y_train, epochs = 100, validation_data = (x_test, y_test))",
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": "Train on 120 samples, validate on 30 samples\nEpoch 1/100\n120/120 [==============================] - 1s 12ms/step - loss: 1.0881 - acc: 0.4417 - val_loss: 1.0510 - val_acc: 0.7000\nEpoch 2/100\n120/120 [==============================] - 0s 292us/step - loss: 1.0290 - acc: 0.6583 - val_loss: 0.9233 - val_acc: 0.7000\nEpoch 3/100\n120/120 [==============================] - 0s 297us/step - loss: 0.8841 - acc: 0.6583 - val_loss: 0.7049 - val_acc: 0.7000\nEpoch 4/100\n120/120 [==============================] - 0s 281us/step - loss: 0.6811 - acc: 0.6583 - val_loss: 0.4937 - val_acc: 0.7000\nEpoch 5/100\n120/120 [==============================] - 0s 287us/step - loss: 0.4879 - acc: 0.7083 - val_loss: 0.3887 - val_acc: 0.8333\nEpoch 6/100\n120/120 [==============================] - 0s 283us/step - loss: 0.4193 - acc: 0.8750 - val_loss: 0.3284 - val_acc: 0.8333\nEpoch 7/100\n120/120 [==============================] - 0s 299us/step - loss: 0.3478 - acc: 0.8250 - val_loss: 0.2704 - val_acc: 0.9000\nEpoch 8/100\n120/120 [==============================] - 0s 306us/step - loss: 0.2975 - acc: 0.8750 - val_loss: 0.2040 - val_acc: 0.9000\nEpoch 9/100\n120/120 [==============================] - 0s 308us/step - loss: 0.2378 - acc: 0.9167 - val_loss: 0.1592 - val_acc: 0.9667\nEpoch 10/100\n120/120 [==============================] - 0s 285us/step - loss: 0.2822 - acc: 0.8833 - val_loss: 0.1440 - val_acc: 0.9000\nEpoch 11/100\n120/120 [==============================] - 0s 296us/step - loss: 0.2005 - acc: 0.9333 - val_loss: 0.0797 - val_acc: 1.0000\nEpoch 12/100\n120/120 [==============================] - 0s 284us/step - loss: 0.1982 - acc: 0.9167 - val_loss: 0.1128 - val_acc: 0.9667\nEpoch 13/100\n120/120 [==============================] - 0s 286us/step - loss: 0.1395 - acc: 0.9500 - val_loss: 0.0565 - val_acc: 1.0000\nEpoch 14/100\n120/120 [==============================] - 0s 288us/step - loss: 0.1515 - acc: 0.9583 - val_loss: 0.0641 - val_acc: 0.9667\nEpoch 15/100\n120/120 [==============================] - 0s 273us/step - loss: 0.1777 - acc: 0.9167 - val_loss: 0.0627 - val_acc: 1.0000\nEpoch 16/100\n120/120 [==============================] - 0s 276us/step - loss: 0.1357 - acc: 0.9583 - val_loss: 0.0401 - val_acc: 1.0000\nEpoch 17/100\n120/120 [==============================] - 0s 297us/step - loss: 0.1208 - acc: 0.9500 - val_loss: 0.0727 - val_acc: 0.9667\nEpoch 18/100\n120/120 [==============================] - 0s 294us/step - loss: 0.1852 - acc: 0.9083 - val_loss: 0.0917 - val_acc: 0.9667\nEpoch 19/100\n120/120 [==============================] - 0s 294us/step - loss: 0.1446 - acc: 0.9417 - val_loss: 0.0499 - val_acc: 1.0000\nEpoch 20/100\n120/120 [==============================] - 0s 289us/step - loss: 0.1457 - acc: 0.9500 - val_loss: 0.0708 - val_acc: 0.9667\nEpoch 21/100\n120/120 [==============================] - 0s 276us/step - loss: 0.1346 - acc: 0.9333 - val_loss: 0.0838 - val_acc: 0.9667\nEpoch 22/100\n120/120 [==============================] - 0s 277us/step - loss: 0.0952 - acc: 0.9667 - val_loss: 0.0341 - val_acc: 1.0000\nEpoch 23/100\n120/120 [==============================] - 0s 286us/step - loss: 0.1370 - acc: 0.9500 - val_loss: 0.0311 - val_acc: 1.0000\nEpoch 24/100\n120/120 [==============================] - 0s 294us/step - loss: 0.0984 - acc: 0.9667 - val_loss: 0.0949 - val_acc: 0.9333\nEpoch 25/100\n120/120 [==============================] - 0s 299us/step - loss: 0.1088 - acc: 0.9417 - val_loss: 0.0506 - val_acc: 0.9667\nEpoch 26/100\n120/120 [==============================] - 0s 301us/step - loss: 0.0840 - acc: 0.9667 - val_loss: 0.0319 - val_acc: 1.0000\nEpoch 27/100\n120/120 [==============================] - 0s 285us/step - loss: 0.0876 - acc: 0.9667 - val_loss: 0.0340 - val_acc: 1.0000\nEpoch 28/100\n120/120 [==============================] - 0s 305us/step - loss: 0.0807 - acc: 0.9667 - val_loss: 0.0532 - val_acc: 0.9667\nEpoch 29/100\n120/120 [==============================] - 0s 291us/step - loss: 0.0778 - acc: 0.9583 - val_loss: 0.0384 - val_acc: 1.0000\nEpoch 30/100\n120/120 [==============================] - 0s 289us/step - loss: 0.0742 - acc: 0.9667 - val_loss: 0.0367 - val_acc: 1.0000\nEpoch 31/100\n120/120 [==============================] - 0s 278us/step - loss: 0.0726 - acc: 0.9667 - val_loss: 0.0378 - val_acc: 0.9667\nEpoch 32/100\n120/120 [==============================] - 0s 284us/step - loss: 0.0735 - acc: 0.9667 - val_loss: 0.0348 - val_acc: 0.9667\nEpoch 33/100\n120/120 [==============================] - 0s 296us/step - loss: 0.0721 - acc: 0.9667 - val_loss: 0.0515 - val_acc: 0.9667\nEpoch 34/100\n120/120 [==============================] - 0s 280us/step - loss: 0.0760 - acc: 0.9583 - val_loss: 0.0321 - val_acc: 1.0000\nEpoch 35/100\n120/120 [==============================] - 0s 289us/step - loss: 0.0684 - acc: 0.9750 - val_loss: 0.0418 - val_acc: 0.9667\nEpoch 36/100\n120/120 [==============================] - 0s 278us/step - loss: 0.0683 - acc: 0.9750 - val_loss: 0.0410 - val_acc: 0.9667\nEpoch 37/100\n120/120 [==============================] - 0s 282us/step - loss: 0.0698 - acc: 0.9667 - val_loss: 0.0425 - val_acc: 0.9667\nEpoch 38/100\n120/120 [==============================] - 0s 293us/step - loss: 0.0691 - acc: 0.9750 - val_loss: 0.0306 - val_acc: 1.0000\nEpoch 39/100\n120/120 [==============================] - 0s 285us/step - loss: 0.0693 - acc: 0.9750 - val_loss: 0.0540 - val_acc: 0.9667\nEpoch 40/100\n120/120 [==============================] - 0s 294us/step - loss: 0.0727 - acc: 0.9750 - val_loss: 0.0322 - val_acc: 1.0000\nEpoch 41/100\n120/120 [==============================] - 0s 286us/step - loss: 0.0648 - acc: 0.9750 - val_loss: 0.0409 - val_acc: 0.9667\nEpoch 42/100\n120/120 [==============================] - 0s 294us/step - loss: 0.0663 - acc: 0.9583 - val_loss: 0.0444 - val_acc: 0.9667\nEpoch 43/100\n120/120 [==============================] - 0s 294us/step - loss: 0.0665 - acc: 0.9667 - val_loss: 0.0443 - val_acc: 0.9667\nEpoch 44/100\n120/120 [==============================] - 0s 297us/step - loss: 0.0619 - acc: 0.9750 - val_loss: 0.0382 - val_acc: 0.9667\nEpoch 45/100\n120/120 [==============================] - 0s 287us/step - loss: 0.0642 - acc: 0.9750 - val_loss: 0.0392 - val_acc: 0.9667\nEpoch 46/100\n120/120 [==============================] - 0s 275us/step - loss: 0.0595 - acc: 0.9750 - val_loss: 0.0454 - val_acc: 0.9667\nEpoch 47/100\n120/120 [==============================] - 0s 279us/step - loss: 0.0611 - acc: 0.9750 - val_loss: 0.0401 - val_acc: 0.9667\nEpoch 48/100\n120/120 [==============================] - 0s 293us/step - loss: 0.0616 - acc: 0.9750 - val_loss: 0.0352 - val_acc: 0.9667\nEpoch 49/100\n120/120 [==============================] - 0s 310us/step - loss: 0.0581 - acc: 0.9750 - val_loss: 0.0495 - val_acc: 0.9667\nEpoch 50/100\n120/120 [==============================] - 0s 309us/step - loss: 0.0600 - acc: 0.9667 - val_loss: 0.0528 - val_acc: 0.9667\nEpoch 51/100\n120/120 [==============================] - 0s 291us/step - loss: 0.0621 - acc: 0.9667 - val_loss: 0.0401 - val_acc: 0.9667\nEpoch 52/100\n120/120 [==============================] - 0s 282us/step - loss: 0.0569 - acc: 0.9667 - val_loss: 0.0480 - val_acc: 0.9667\nEpoch 53/100\n120/120 [==============================] - 0s 279us/step - loss: 0.0587 - acc: 0.9667 - val_loss: 0.0400 - val_acc: 0.9667\nEpoch 54/100\n120/120 [==============================] - 0s 340us/step - loss: 0.0574 - acc: 0.9750 - val_loss: 0.0447 - val_acc: 0.9667\nEpoch 55/100\n120/120 [==============================] - 0s 285us/step - loss: 0.0629 - acc: 0.9583 - val_loss: 0.0445 - val_acc: 0.9667\nEpoch 56/100\n120/120 [==============================] - 0s 298us/step - loss: 0.0585 - acc: 0.9667 - val_loss: 0.0360 - val_acc: 0.9667\nEpoch 57/100\n120/120 [==============================] - 0s 275us/step - loss: 0.0661 - acc: 0.9667 - val_loss: 0.0608 - val_acc: 0.9667\nEpoch 58/100\n120/120 [==============================] - 0s 292us/step - loss: 0.0813 - acc: 0.9500 - val_loss: 0.0265 - val_acc: 1.0000\nEpoch 59/100\n120/120 [==============================] - 0s 282us/step - loss: 0.0554 - acc: 0.9750 - val_loss: 0.0597 - val_acc: 0.9667\nEpoch 60/100\n120/120 [==============================] - 0s 285us/step - loss: 0.0715 - acc: 0.9583 - val_loss: 0.0711 - val_acc: 0.9667\n",
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": "Epoch 61/100\n120/120 [==============================] - 0s 289us/step - loss: 0.0645 - acc: 0.9583 - val_loss: 0.0343 - val_acc: 0.9667\nEpoch 62/100\n120/120 [==============================] - 0s 295us/step - loss: 0.0810 - acc: 0.9750 - val_loss: 0.0418 - val_acc: 0.9667\nEpoch 63/100\n120/120 [==============================] - 0s 277us/step - loss: 0.0560 - acc: 0.9583 - val_loss: 0.0790 - val_acc: 0.9667\nEpoch 64/100\n120/120 [==============================] - 0s 287us/step - loss: 0.0703 - acc: 0.9667 - val_loss: 0.0478 - val_acc: 0.9667\nEpoch 65/100\n120/120 [==============================] - 0s 301us/step - loss: 0.0731 - acc: 0.9667 - val_loss: 0.0272 - val_acc: 1.0000\nEpoch 66/100\n120/120 [==============================] - 0s 301us/step - loss: 0.0647 - acc: 0.9667 - val_loss: 0.0710 - val_acc: 0.9667\nEpoch 67/100\n120/120 [==============================] - 0s 307us/step - loss: 0.0698 - acc: 0.9583 - val_loss: 0.0587 - val_acc: 0.9667\nEpoch 68/100\n120/120 [==============================] - 0s 297us/step - loss: 0.0586 - acc: 0.9667 - val_loss: 0.0297 - val_acc: 1.0000\nEpoch 69/100\n120/120 [==============================] - 0s 282us/step - loss: 0.0688 - acc: 0.9750 - val_loss: 0.0396 - val_acc: 0.9667\nEpoch 70/100\n120/120 [==============================] - 0s 299us/step - loss: 0.0518 - acc: 0.9750 - val_loss: 0.0745 - val_acc: 0.9667\nEpoch 71/100\n120/120 [==============================] - 0s 277us/step - loss: 0.0658 - acc: 0.9583 - val_loss: 0.0449 - val_acc: 0.9667\nEpoch 72/100\n120/120 [==============================] - 0s 285us/step - loss: 0.0524 - acc: 0.9750 - val_loss: 0.0273 - val_acc: 1.0000\nEpoch 73/100\n120/120 [==============================] - 0s 288us/step - loss: 0.0604 - acc: 0.9750 - val_loss: 0.0387 - val_acc: 0.9667\nEpoch 74/100\n120/120 [==============================] - 0s 292us/step - loss: 0.0541 - acc: 0.9583 - val_loss: 0.0587 - val_acc: 0.9667\nEpoch 75/100\n120/120 [==============================] - 0s 289us/step - loss: 0.0572 - acc: 0.9583 - val_loss: 0.0423 - val_acc: 0.9667\nEpoch 76/100\n120/120 [==============================] - 0s 299us/step - loss: 0.0515 - acc: 0.9750 - val_loss: 0.0265 - val_acc: 0.9667\nEpoch 77/100\n120/120 [==============================] - 0s 307us/step - loss: 0.0595 - acc: 0.9667 - val_loss: 0.0413 - val_acc: 0.9667\nEpoch 78/100\n120/120 [==============================] - 0s 302us/step - loss: 0.0716 - acc: 0.9583 - val_loss: 0.0741 - val_acc: 0.9667\nEpoch 79/100\n120/120 [==============================] - 0s 285us/step - loss: 0.0581 - acc: 0.9667 - val_loss: 0.0347 - val_acc: 0.9667\nEpoch 80/100\n120/120 [==============================] - 0s 294us/step - loss: 0.0607 - acc: 0.9833 - val_loss: 0.0399 - val_acc: 0.9667\nEpoch 81/100\n120/120 [==============================] - 0s 297us/step - loss: 0.0566 - acc: 0.9667 - val_loss: 0.0644 - val_acc: 0.9667\nEpoch 82/100\n120/120 [==============================] - 0s 309us/step - loss: 0.0591 - acc: 0.9583 - val_loss: 0.0402 - val_acc: 0.9667\nEpoch 83/100\n120/120 [==============================] - 0s 290us/step - loss: 0.0531 - acc: 0.9750 - val_loss: 0.0424 - val_acc: 0.9667\nEpoch 84/100\n120/120 [==============================] - 0s 302us/step - loss: 0.0519 - acc: 0.9750 - val_loss: 0.0335 - val_acc: 0.9667\nEpoch 85/100\n120/120 [==============================] - 0s 307us/step - loss: 0.0520 - acc: 0.9750 - val_loss: 0.0382 - val_acc: 0.9667\nEpoch 86/100\n120/120 [==============================] - 0s 290us/step - loss: 0.0507 - acc: 0.9750 - val_loss: 0.0571 - val_acc: 0.9667\nEpoch 87/100\n120/120 [==============================] - 0s 293us/step - loss: 0.0586 - acc: 0.9667 - val_loss: 0.0418 - val_acc: 0.9667\nEpoch 88/100\n120/120 [==============================] - 0s 291us/step - loss: 0.0612 - acc: 0.9583 - val_loss: 0.0698 - val_acc: 0.9667\nEpoch 89/100\n120/120 [==============================] - 0s 291us/step - loss: 0.0536 - acc: 0.9667 - val_loss: 0.0311 - val_acc: 0.9667\nEpoch 90/100\n120/120 [==============================] - 0s 295us/step - loss: 0.0520 - acc: 0.9750 - val_loss: 0.0317 - val_acc: 0.9667\nEpoch 91/100\n120/120 [==============================] - 0s 296us/step - loss: 0.0531 - acc: 0.9667 - val_loss: 0.0482 - val_acc: 0.9667\nEpoch 92/100\n120/120 [==============================] - 0s 288us/step - loss: 0.0502 - acc: 0.9667 - val_loss: 0.0332 - val_acc: 0.9667\nEpoch 93/100\n120/120 [==============================] - 0s 296us/step - loss: 0.0471 - acc: 0.9750 - val_loss: 0.0453 - val_acc: 0.9667\nEpoch 94/100\n120/120 [==============================] - 0s 291us/step - loss: 0.0447 - acc: 0.9667 - val_loss: 0.0667 - val_acc: 0.9667\nEpoch 95/100\n120/120 [==============================] - 0s 283us/step - loss: 0.0506 - acc: 0.9667 - val_loss: 0.0366 - val_acc: 0.9667\nEpoch 96/100\n120/120 [==============================] - 0s 281us/step - loss: 0.0466 - acc: 0.9667 - val_loss: 0.0577 - val_acc: 0.9667\nEpoch 97/100\n120/120 [==============================] - 0s 281us/step - loss: 0.0443 - acc: 0.9667 - val_loss: 0.0366 - val_acc: 0.9667\nEpoch 98/100\n120/120 [==============================] - 0s 280us/step - loss: 0.0467 - acc: 0.9667 - val_loss: 0.1003 - val_acc: 0.9667\nEpoch 99/100\n120/120 [==============================] - 0s 293us/step - loss: 0.0576 - acc: 0.9667 - val_loss: 0.0455 - val_acc: 0.9667\nEpoch 100/100\n120/120 [==============================] - 0s 288us/step - loss: 0.0537 - acc: 0.9833 - val_loss: 0.0290 - val_acc: 1.0000\n",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "700989a441f89101245ae099ebe528fe3f633431"
      },
      "cell_type": "code",
      "source": "plt.plot(deep_history.history['acc'])\nplt.plot(deep_history.history['val_acc'])\nplt.title(\"Accuracy\")\nplt.legend(['train', 'test'])\nplt.show()",
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "e11dd866567ce68bd5b4050d06793d33ae78a622"
      },
      "cell_type": "code",
      "source": "plt.plot(deep_history.history['loss'])\nplt.plot(deep_history.history['val_loss'])\nplt.title('Loss')\nplt.legend(['Train','Test'])\nplt.show()",
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "_uuid": "c7bc4c3d21891be9ebed38cdd359b1963cf6d5aa"
      },
      "cell_type": "markdown",
      "source": "So our deep model is more accurate than the other models."
    },
    {
      "metadata": {
        "trusted": true,
        "_uuid": "c31866c872647caccbc1557a0e1ef39688c38cf6"
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "name": "python",
      "version": "3.6.6",
      "mimetype": "text/x-python",
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "pygments_lexer": "ipython3",
      "nbconvert_exporter": "python",
      "file_extension": ".py"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 1
}