SVM, Decision Tree, CNN on TORONTO dataset.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "SER USING TORONTO SPEECH DATASET.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ShireenPatel/speech-emotion-recognition/blob/master/SVM%2C%20Decision%20Tree%2C%20CNN%20on%20TORONTO%20dataset.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lDdJRnowmW9U",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!pip install tensorflow==1.14.0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f3BSkSCkgHpJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "1147b43f-9688-49ce-8f18-beafb789211b"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SvBkohPDgYUk",
        "colab_type": "code",
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "Ly8gQ29weXJpZ2h0IDIwMTcgR29vZ2xlIExMQwovLwovLyBMaWNlbnNlZCB1bmRlciB0aGUgQXBhY2hlIExpY2Vuc2UsIFZlcnNpb24gMi4wICh0aGUgIkxpY2Vuc2UiKTsKLy8geW91IG1heSBub3QgdXNlIHRoaXMgZmlsZSBleGNlcHQgaW4gY29tcGxpYW5jZSB3aXRoIHRoZSBMaWNlbnNlLgovLyBZb3UgbWF5IG9idGFpbiBhIGNvcHkgb2YgdGhlIExpY2Vuc2UgYXQKLy8KLy8gICAgICBodHRwOi8vd3d3LmFwYWNoZS5vcmcvbGljZW5zZXMvTElDRU5TRS0yLjAKLy8KLy8gVW5sZXNzIHJlcXVpcmVkIGJ5IGFwcGxpY2FibGUgbGF3IG9yIGFncmVlZCB0byBpbiB3cml0aW5nLCBzb2Z0d2FyZQovLyBkaXN0cmlidXRlZCB1bmRlciB0aGUgTGljZW5zZSBpcyBkaXN0cmlidXRlZCBvbiBhbiAiQVMgSVMiIEJBU0lTLAovLyBXSVRIT1VUIFdBUlJBTlRJRVMgT1IgQ09ORElUSU9OUyBPRiBBTlkgS0lORCwgZWl0aGVyIGV4cHJlc3Mgb3IgaW1wbGllZC4KLy8gU2VlIHRoZSBMaWNlbnNlIGZvciB0aGUgc3BlY2lmaWMgbGFuZ3VhZ2UgZ292ZXJuaW5nIHBlcm1pc3Npb25zIGFuZAovLyBsaW1pdGF0aW9ucyB1bmRlciB0aGUgTGljZW5zZS4KCi8qKgogKiBAZmlsZW92ZXJ2aWV3IEhlbHBlcnMgZm9yIGdvb2dsZS5jb2xhYiBQeXRob24gbW9kdWxlLgogKi8KKGZ1bmN0aW9uKHNjb3BlKSB7CmZ1bmN0aW9uIHNwYW4odGV4dCwgc3R5bGVBdHRyaWJ1dGVzID0ge30pIHsKICBjb25zdCBlbGVtZW50ID0gZG9jdW1lbnQuY3JlYXRlRWxlbWVudCgnc3BhbicpOwogIGVsZW1lbnQudGV4dENvbnRlbnQgPSB0ZXh0OwogIGZvciAoY29uc3Qga2V5IG9mIE9iamVjdC5rZXlzKHN0eWxlQXR0cmlidXRlcykpIHsKICAgIGVsZW1lbnQuc3R5bGVba2V5XSA9IHN0eWxlQXR0cmlidXRlc1trZXldOwogIH0KICByZXR1cm4gZWxlbWVudDsKfQoKLy8gTWF4IG51bWJlciBvZiBieXRlcyB3aGljaCB3aWxsIGJlIHVwbG9hZGVkIGF0IGEgdGltZS4KY29uc3QgTUFYX1BBWUxPQURfU0laRSA9IDEwMCAqIDEwMjQ7Ci8vIE1heCBhbW91bnQgb2YgdGltZSB0byBibG9jayB3YWl0aW5nIGZvciB0aGUgdXNlci4KY29uc3QgRklMRV9DSEFOR0VfVElNRU9VVF9NUyA9IDMwICogMTAwMDsKCmZ1bmN0aW9uIF91cGxvYWRGaWxlcyhpbnB1dElkLCBvdXRwdXRJZCkgewogIGNvbnN0IHN0ZXBzID0gdXBsb2FkRmlsZXNTdGVwKGlucHV0SWQsIG91dHB1dElkKTsKICBjb25zdCBvdXRwdXRFbGVtZW50ID0gZG9jdW1lbnQuZ2V0RWxlbWVudEJ5SWQob3V0cHV0SWQpOwogIC8vIENhY2hlIHN0ZXBzIG9uIHRoZSBvdXRwdXRFbGVtZW50IHRvIG1ha2UgaXQgYXZhaWxhYmxlIGZvciB0aGUgbmV4dCBjYWxsCiAgLy8gdG8gdXBsb2FkRmlsZXNDb250aW51ZSBmcm9tIFB5dGhvbi4KICBvdXRwdXRFbGVtZW50LnN0ZXBzID0gc3RlcHM7CgogIHJldHVybiBfdXBsb2FkRmlsZXNDb250aW51ZShvdXRwdXRJZCk7Cn0KCi8vIFRoaXMgaXMgcm91Z2hseSBhbiBhc3luYyBnZW5lcmF0b3IgKG5vdCBzdXBwb3J0ZWQgaW4gdGhlIGJyb3dzZXIgeWV0KSwKLy8gd2hlcmUgdGhlcmUgYXJlIG11bHRpcGxlIGFzeW5jaHJvbm91cyBzdGVwcyBhbmQgdGhlIFB5dGhvbiBzaWRlIGlzIGdvaW5nCi8vIHRvIHBvbGwgZm9yIGNvbXBsZXRpb24gb2YgZWFjaCBzdGVwLgovLyBUaGlzIHVzZXMgYSBQcm9taXNlIHRvIGJsb2NrIHRoZSBweXRob24gc2lkZSBvbiBjb21wbGV0aW9uIG9mIGVhY2ggc3RlcCwKLy8gdGhlbiBwYXNzZXMgdGhlIHJlc3VsdCBvZiB0aGUgcHJldmlvdXMgc3RlcCBhcyB0aGUgaW5wdXQgdG8gdGhlIG5leHQgc3RlcC4KZnVuY3Rpb24gX3VwbG9hZEZpbGVzQ29udGludWUob3V0cHV0SWQpIHsKICBjb25zdCBvdXRwdXRFbGVtZW50ID0gZG9jdW1lbnQuZ2V0RWxlbWVudEJ5SWQob3V0cHV0SWQpOwogIGNvbnN0IHN0ZXBzID0gb3V0cHV0RWxlbWVudC5zdGVwczsKCiAgY29uc3QgbmV4dCA9IHN0ZXBzLm5leHQob3V0cHV0RWxlbWVudC5sYXN0UHJvbWlzZVZhbHVlKTsKICByZXR1cm4gUHJvbWlzZS5yZXNvbHZlKG5leHQudmFsdWUucHJvbWlzZSkudGhlbigodmFsdWUpID0+IHsKICAgIC8vIENhY2hlIHRoZSBsYXN0IHByb21pc2UgdmFsdWUgdG8gbWFrZSBpdCBhdmFpbGFibGUgdG8gdGhlIG5leHQKICAgIC8vIHN0ZXAgb2YgdGhlIGdlbmVyYXRvci4KICAgIG91dHB1dEVsZW1lbnQubGFzdFByb21pc2VWYWx1ZSA9IHZhbHVlOwogICAgcmV0dXJuIG5leHQudmFsdWUucmVzcG9uc2U7CiAgfSk7Cn0KCi8qKgogKiBHZW5lcmF0b3IgZnVuY3Rpb24gd2hpY2ggaXMgY2FsbGVkIGJldHdlZW4gZWFjaCBhc3luYyBzdGVwIG9mIHRoZSB1cGxvYWQKICogcHJvY2Vzcy4KICogQHBhcmFtIHtzdHJpbmd9IGlucHV0SWQgRWxlbWVudCBJRCBvZiB0aGUgaW5wdXQgZmlsZSBwaWNrZXIgZWxlbWVudC4KICogQHBhcmFtIHtzdHJpbmd9IG91dHB1dElkIEVsZW1lbnQgSUQgb2YgdGhlIG91dHB1dCBkaXNwbGF5LgogKiBAcmV0dXJuIHshSXRlcmFibGU8IU9iamVjdD59IEl0ZXJhYmxlIG9mIG5leHQgc3RlcHMuCiAqLwpmdW5jdGlvbiogdXBsb2FkRmlsZXNTdGVwKGlucHV0SWQsIG91dHB1dElkKSB7CiAgY29uc3QgaW5wdXRFbGVtZW50ID0gZG9jdW1lbnQuZ2V0RWxlbWVudEJ5SWQoaW5wdXRJZCk7CiAgaW5wdXRFbGVtZW50LmRpc2FibGVkID0gZmFsc2U7CgogIGNvbnN0IG91dHB1dEVsZW1lbnQgPSBkb2N1bWVudC5nZXRFbGVtZW50QnlJZChvdXRwdXRJZCk7CiAgb3V0cHV0RWxlbWVudC5pbm5lckhUTUwgPSAnJzsKCiAgY29uc3QgcGlja2VkUHJvbWlzZSA9IG5ldyBQcm9taXNlKChyZXNvbHZlKSA9PiB7CiAgICBpbnB1dEVsZW1lbnQuYWRkRXZlbnRMaXN0ZW5lcignY2hhbmdlJywgKGUpID0+IHsKICAgICAgcmVzb2x2ZShlLnRhcmdldC5maWxlcyk7CiAgICB9KTsKICB9KTsKCiAgY29uc3QgY2FuY2VsID0gZG9jdW1lbnQuY3JlYXRlRWxlbWVudCgnYnV0dG9uJyk7CiAgaW5wdXRFbGVtZW50LnBhcmVudEVsZW1lbnQuYXBwZW5kQ2hpbGQoY2FuY2VsKTsKICBjYW5jZWwudGV4dENvbnRlbnQgPSAnQ2FuY2VsIHVwbG9hZCc7CiAgY29uc3QgY2FuY2VsUHJvbWlzZSA9IG5ldyBQcm9taXNlKChyZXNvbHZlKSA9PiB7CiAgICBjYW5jZWwub25jbGljayA9ICgpID0+IHsKICAgICAgcmVzb2x2ZShudWxsKTsKICAgIH07CiAgfSk7CgogIC8vIENhbmNlbCB1cGxvYWQgaWYgdXNlciBoYXNuJ3QgcGlja2VkIGFueXRoaW5nIGluIHRpbWVvdXQuCiAgY29uc3QgdGltZW91dFByb21pc2UgPSBuZXcgUHJvbWlzZSgocmVzb2x2ZSkgPT4gewogICAgc2V0VGltZW91dCgoKSA9PiB7CiAgICAgIHJlc29sdmUobnVsbCk7CiAgICB9LCBGSUxFX0NIQU5HRV9USU1FT1VUX01TKTsKICB9KTsKCiAgLy8gV2FpdCBmb3IgdGhlIHVzZXIgdG8gcGljayB0aGUgZmlsZXMuCiAgY29uc3QgZmlsZXMgPSB5aWVsZCB7CiAgICBwcm9taXNlOiBQcm9taXNlLnJhY2UoW3BpY2tlZFByb21pc2UsIHRpbWVvdXRQcm9taXNlLCBjYW5jZWxQcm9taXNlXSksCiAgICByZXNwb25zZTogewogICAgICBhY3Rpb246ICdzdGFydGluZycsCiAgICB9CiAgfTsKCiAgaWYgKCFmaWxlcykgewogICAgcmV0dXJuIHsKICAgICAgcmVzcG9uc2U6IHsKICAgICAgICBhY3Rpb246ICdjb21wbGV0ZScsCiAgICAgIH0KICAgIH07CiAgfQoKICBjYW5jZWwucmVtb3ZlKCk7CgogIC8vIERpc2FibGUgdGhlIGlucHV0IGVsZW1lbnQgc2luY2UgZnVydGhlciBwaWNrcyBhcmUgbm90IGFsbG93ZWQuCiAgaW5wdXRFbGVtZW50LmRpc2FibGVkID0gdHJ1ZTsKCiAgZm9yIChjb25zdCBmaWxlIG9mIGZpbGVzKSB7CiAgICBjb25zdCBsaSA9IGRvY3VtZW50LmNyZWF0ZUVsZW1lbnQoJ2xpJyk7CiAgICBsaS5hcHBlbmQoc3BhbihmaWxlLm5hbWUsIHtmb250V2VpZ2h0OiAnYm9sZCd9KSk7CiAgICBsaS5hcHBlbmQoc3BhbigKICAgICAgICBgKCR7ZmlsZS50eXBlIHx8ICduL2EnfSkgLSAke2ZpbGUuc2l6ZX0gYnl0ZXMsIGAgKwogICAgICAgIGBsYXN0IG1vZGlmaWVkOiAkewogICAgICAgICAgICBmaWxlLmxhc3RNb2RpZmllZERhdGUgPyBmaWxlLmxhc3RNb2RpZmllZERhdGUudG9Mb2NhbGVEYXRlU3RyaW5nKCkgOgogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAnbi9hJ30gLSBgKSk7CiAgICBjb25zdCBwZXJjZW50ID0gc3BhbignMCUgZG9uZScpOwogICAgbGkuYXBwZW5kQ2hpbGQocGVyY2VudCk7CgogICAgb3V0cHV0RWxlbWVudC5hcHBlbmRDaGlsZChsaSk7CgogICAgY29uc3QgZmlsZURhdGFQcm9taXNlID0gbmV3IFByb21pc2UoKHJlc29sdmUpID0+IHsKICAgICAgY29uc3QgcmVhZGVyID0gbmV3IEZpbGVSZWFkZXIoKTsKICAgICAgcmVhZGVyLm9ubG9hZCA9IChlKSA9PiB7CiAgICAgICAgcmVzb2x2ZShlLnRhcmdldC5yZXN1bHQpOwogICAgICB9OwogICAgICByZWFkZXIucmVhZEFzQXJyYXlCdWZmZXIoZmlsZSk7CiAgICB9KTsKICAgIC8vIFdhaXQgZm9yIHRoZSBkYXRhIHRvIGJlIHJlYWR5LgogICAgbGV0IGZpbGVEYXRhID0geWllbGQgewogICAgICBwcm9taXNlOiBmaWxlRGF0YVByb21pc2UsCiAgICAgIHJlc3BvbnNlOiB7CiAgICAgICAgYWN0aW9uOiAnY29udGludWUnLAogICAgICB9CiAgICB9OwoKICAgIC8vIFVzZSBhIGNodW5rZWQgc2VuZGluZyB0byBhdm9pZCBtZXNzYWdlIHNpemUgbGltaXRzLiBTZWUgYi82MjExNTY2MC4KICAgIGxldCBwb3NpdGlvbiA9IDA7CiAgICB3aGlsZSAocG9zaXRpb24gPCBmaWxlRGF0YS5ieXRlTGVuZ3RoKSB7CiAgICAgIGNvbnN0IGxlbmd0aCA9IE1hdGgubWluKGZpbGVEYXRhLmJ5dGVMZW5ndGggLSBwb3NpdGlvbiwgTUFYX1BBWUxPQURfU0laRSk7CiAgICAgIGNvbnN0IGNodW5rID0gbmV3IFVpbnQ4QXJyYXkoZmlsZURhdGEsIHBvc2l0aW9uLCBsZW5ndGgpOwogICAgICBwb3NpdGlvbiArPSBsZW5ndGg7CgogICAgICBjb25zdCBiYXNlNjQgPSBidG9hKFN0cmluZy5mcm9tQ2hhckNvZGUuYXBwbHkobnVsbCwgY2h1bmspKTsKICAgICAgeWllbGQgewogICAgICAgIHJlc3BvbnNlOiB7CiAgICAgICAgICBhY3Rpb246ICdhcHBlbmQnLAogICAgICAgICAgZmlsZTogZmlsZS5uYW1lLAogICAgICAgICAgZGF0YTogYmFzZTY0LAogICAgICAgIH0sCiAgICAgIH07CiAgICAgIHBlcmNlbnQudGV4dENvbnRlbnQgPQogICAgICAgICAgYCR7TWF0aC5yb3VuZCgocG9zaXRpb24gLyBmaWxlRGF0YS5ieXRlTGVuZ3RoKSAqIDEwMCl9JSBkb25lYDsKICAgIH0KICB9CgogIC8vIEFsbCBkb25lLgogIHlpZWxkIHsKICAgIHJlc3BvbnNlOiB7CiAgICAgIGFjdGlvbjogJ2NvbXBsZXRlJywKICAgIH0KICB9Owp9CgpzY29wZS5nb29nbGUgPSBzY29wZS5nb29nbGUgfHwge307CnNjb3BlLmdvb2dsZS5jb2xhYiA9IHNjb3BlLmdvb2dsZS5jb2xhYiB8fCB7fTsKc2NvcGUuZ29vZ2xlLmNvbGFiLl9maWxlcyA9IHsKICBfdXBsb2FkRmlsZXMsCiAgX3VwbG9hZEZpbGVzQ29udGludWUsCn07Cn0pKHNlbGYpOwo=",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 91
        },
        "outputId": "b1c1af36-0b5e-4aab-f843-39e2dac00a06"
      },
      "source": [
        "from google.colab import files\n",
        "!pip install -q kaggle\n",
        "from google.colab import files\n",
        "files.upload()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-d6b32966-eef6-418a-87c8-0c2cb57600a5\" name=\"files[]\" multiple disabled />\n",
              "     <output id=\"result-d6b32966-eef6-418a-87c8-0c2cb57600a5\">\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 kaggle.json to kaggle (1).json\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'kaggle.json': b'{\"username\":\"patelshireendharmesh\",\"key\":\"a383b88dc708b648ed3d3741fbd96af0\"}'}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dWzGhQj7gfxl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!mkdir -p ~/.kaggle\n",
        "!cp kaggle.json ~/.kaggle/\n",
        "!chmod 600 ~/.kaggle/kaggle.json"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g_GE4gluglnT",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "6b7441d7-004d-40db-e351-c321920f1132"
      },
      "source": [
        "!kaggle datasets download -d ejlok1/toronto-emotional-speech-set-tess"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "toronto-emotional-speech-set-tess.zip: Skipping, found more recently modified local copy (use --force to force download)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2lzZu-3TgskG",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "outputId": "e73c3c8a-b223-4ff3-a114-15a1f4c0ac15"
      },
      "source": [
        "!ls"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            " drive\t\t    kaggle.json   toronto\n",
            "'kaggle (1).json'   sample_data   toronto-emotional-speech-set-tess.zip\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YGOqF36PgwdH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        },
        "outputId": "31439e23-8fcb-4f69-f627-083315bfd7b6"
      },
      "source": [
        "!mkdir toronto\n",
        "!unzip toronto-emotional-speech-set-tess.zip -d toronto"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "mkdir: cannot create directory ‘toronto’: File exists\n",
            "Archive:  toronto-emotional-speech-set-tess.zip\n",
            "replace toronto/TESS Toronto emotional speech set data/OAF_Fear/OAF_back_fear.wav? [y]es, [n]o, [A]ll, [N]one, [r]ename: "
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SugMvD4ug4k8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 218
        },
        "outputId": "2e8c776b-7ebf-4d87-b30d-d3c527cd3d87"
      },
      "source": [
        "!pip install librosa\n",
        "import librosa\n",
        "from librosa import display\n",
        "import os\n",
        "import pandas as pd\n",
        "import glob "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: librosa in /usr/local/lib/python3.6/dist-packages (0.6.3)\n",
            "Requirement already satisfied: audioread>=2.0.0 in /usr/local/lib/python3.6/dist-packages (from librosa) (2.1.8)\n",
            "Requirement already satisfied: joblib>=0.12 in /usr/local/lib/python3.6/dist-packages (from librosa) (0.14.1)\n",
            "Requirement already satisfied: numba>=0.38.0 in /usr/local/lib/python3.6/dist-packages (from librosa) (0.48.0)\n",
            "Requirement already satisfied: scikit-learn!=0.19.0,>=0.14.0 in /usr/local/lib/python3.6/dist-packages (from librosa) (0.22.2.post1)\n",
            "Requirement already satisfied: scipy>=1.0.0 in /usr/local/lib/python3.6/dist-packages (from librosa) (1.4.1)\n",
            "Requirement already satisfied: resampy>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from librosa) (0.2.2)\n",
            "Requirement already satisfied: six>=1.3 in /usr/local/lib/python3.6/dist-packages (from librosa) (1.12.0)\n",
            "Requirement already satisfied: numpy>=1.8.0 in /usr/local/lib/python3.6/dist-packages (from librosa) (1.18.3)\n",
            "Requirement already satisfied: decorator>=3.0.0 in /usr/local/lib/python3.6/dist-packages (from librosa) (4.4.2)\n",
            "Requirement already satisfied: llvmlite<0.32.0,>=0.31.0dev0 in /usr/local/lib/python3.6/dist-packages (from numba>=0.38.0->librosa) (0.31.0)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from numba>=0.38.0->librosa) (46.1.3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CZSaxCB0iP8R",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 252
        },
        "outputId": "e743cc03-1647-4627-ef08-ed34a0e86ba9"
      },
      "source": [
        "dir_list = os.listdir('/content/toronto/TESS Toronto emotional speech set data')\n",
        "dir_list.sort()\n",
        "dir_list"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['OAF_Fear',\n",
              " 'OAF_Pleasant_surprise',\n",
              " 'OAF_Sad',\n",
              " 'OAF_angry',\n",
              " 'OAF_disgust',\n",
              " 'OAF_happy',\n",
              " 'OAF_neutral',\n",
              " 'YAF_angry',\n",
              " 'YAF_disgust',\n",
              " 'YAF_fear',\n",
              " 'YAF_happy',\n",
              " 'YAF_neutral',\n",
              " 'YAF_pleasant_surprised',\n",
              " 'YAF_sad']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yIPJPM14pNIj",
        "colab_type": "text"
      },
      "source": [
        "# New Section"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ANqUmosdlrtl",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "660041dc-f218-4bd8-bdcf-3680751f2d01"
      },
      "source": [
        "import time\n",
        "import numpy as np\n",
        "\n",
        "path = '/content/toronto/TESS Toronto emotional speech set data/'\n",
        "lst = []\n",
        "\n",
        "start_time = time.time()  \n",
        "\n",
        "for subdir, dirs, files in os.walk(path):\n",
        "  for file in files:\n",
        "      try:\n",
        "        #Load librosa array, obtain mfcss, store the file and the mcss information in a new array\n",
        "        X, sample_rate = librosa.load(os.path.join(subdir,file), res_type='kaiser_fast')\n",
        "        mfccs = np.mean(librosa.feature.mfcc(y=X, sr=sample_rate, n_mfcc=200).T,axis=0) \n",
        "        # The instruction below converts the labels (from 1 to 8) to a series from 0 to 7\n",
        "        # This is because our predictor needs to start from 0 otherwise it will try to predict also 0.\n",
        "        #file = int(file[7:8]) \n",
        "        arr = mfccs\n",
        "        lst.append(arr)\n",
        "      # If the file is not valid, skip it\n",
        "      except ValueError:\n",
        "        continue\n",
        "\n",
        "print(\"--- Data loaded. Loading time: %s seconds ---\" % (time.time() - start_time))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--- Data loaded. Loading time: 372.0375256538391 seconds ---\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cd9TyBgqvCKv",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 185
        },
        "outputId": "24574b6b-aa5e-424b-d470-a1f86442a11f"
      },
      "source": [
        "print(lst)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "IOPub data rate exceeded.\n",
            "The notebook server will temporarily stop sending output\n",
            "to the client in order to avoid crashing it.\n",
            "To change this limit, set the config variable\n",
            "`--NotebookApp.iopub_data_rate_limit`.\n",
            "\n",
            "Current values:\n",
            "NotebookApp.iopub_data_rate_limit=1000000.0 (bytes/sec)\n",
            "NotebookApp.rate_limit_window=3.0 (secs)\n",
            "\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5pKxCtAqAHMd",
        "colab_type": "text"
      },
      "source": [
        "Labelling the wav files and storing the numbers in list y :\n",
        "0 = angry, 1 = disgust, 2 = fear, 3 = happy, 4 = neutral, 5 = surprise, 6 = sad"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gf0uYPbrqxRt",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 151
        },
        "outputId": "4eaa25b7-eb96-4a33-d3ea-fdc9470f2af3"
      },
      "source": [
        "TESS= '/content/toronto/TESS Toronto emotional speech set data/'\n",
        "path = []\n",
        "emotion = []\n",
        "y=[]\n",
        "\n",
        "for i in dir_list:\n",
        "    fname = os.listdir('/content/toronto/TESS Toronto emotional speech set data/' + i)\n",
        "    for f in fname:\n",
        "        if i == 'OAF_angry' or i == 'YAF_angry':\n",
        "            emotion.append('female_angry')\n",
        "            y.append(0)\n",
        "        elif i == 'OAF_disgust' or i == 'YAF_disgust':\n",
        "            emotion.append('female_disgust')\n",
        "            y.append(1)\n",
        "        elif i == 'OAF_Fear' or i == 'YAF_fear':\n",
        "            emotion.append('female_fear')\n",
        "            y.append(2)\n",
        "        elif i == 'OAF_happy' or i == 'YAF_happy':\n",
        "            emotion.append('female_happy')\n",
        "            y.append(3)\n",
        "        elif i == 'OAF_neutral' or i == 'YAF_neutral':\n",
        "            emotion.append('female_neutral')\n",
        "            y.append(4)                                \n",
        "        elif i == 'OAF_Pleasant_surprise' or i == 'YAF_pleasant_surprised':\n",
        "            emotion.append('female_surprise') \n",
        "            y.append(5)              \n",
        "        elif i == 'OAF_Sad' or i == 'YAF_sad':\n",
        "            emotion.append('female_sad')\n",
        "            y.append(6)\n",
        "        else:\n",
        "            emotion.append('Unknown')\n",
        "        path.append(TESS + i + \"/\" + f)\n",
        "\n",
        "TESS_df = pd.DataFrame(emotion, columns = ['labels'])\n",
        "TESS_df['source'] = 'TESS'\n",
        "TESS_df = pd.concat([TESS_df,pd.DataFrame(path, columns = ['path'])],axis=1)\n",
        "TESS_df.labels.value_counts()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "female_happy       400\n",
              "female_sad         400\n",
              "female_neutral     400\n",
              "female_surprise    400\n",
              "female_angry       400\n",
              "female_disgust     400\n",
              "female_fear        400\n",
              "Name: labels, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tWyJ33IoUcvm",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "6d2c280f-31a3-470a-de4b-c3f550f93a99"
      },
      "source": [
        "print(y)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ppXDGq3TUkFy",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "7a921b6c-c5b9-4e6d-f3de-0345d4860c82"
      },
      "source": [
        "len(y)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2800"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O20VMaX0wmhP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "61cfb328-48f9-4d4b-899c-08bae9f48648"
      },
      "source": [
        "len(lst)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2800"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "piS_7nBTNxS0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "lst =np.asarray(lst)\n",
        "y=np.asarray(y)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "amI06_Kx8-JW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(lst, y, test_size=0.33, random_state=42)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0rLeoGU09v1t",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "47b6922e-3140-432f-d510-a030d04c0dec"
      },
      "source": [
        "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((1876, 128), (924, 128), (1876,), (924,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FKJL_IvjBGUU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        },
        "outputId": "120d2fbb-2585-4009-c6ee-6dd783faa954"
      },
      "source": [
        "X_train.astype, X_test.astype, y_train.astype, y_test.astype"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<function ndarray.astype>,\n",
              " <function ndarray.astype>,\n",
              " <function ndarray.astype>,\n",
              " <function ndarray.astype>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 104
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "q16X3IGZ_7jc",
        "colab_type": "text"
      },
      "source": [
        "DESICION TREE CLASSIFIER"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1hEAIAG0_DQ2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 252
        },
        "outputId": "3dc441bc-1e11-470d-dbc4-3156352d610d"
      },
      "source": [
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.metrics import accuracy_score\n",
        "\n",
        "dtree = DecisionTreeClassifier()\n",
        "\n",
        "dtree.fit(X_train, y_train)\n",
        "predictions = dtree.predict(X_test)\n",
        "from sklearn.metrics import classification_report,confusion_matrix\n",
        "print(classification_report(y_test,predictions))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.92      0.91      0.91       128\n",
            "           1       0.94      0.91      0.93       125\n",
            "           2       0.93      0.90      0.91       130\n",
            "           3       0.98      0.91      0.94       137\n",
            "           4       0.86      0.93      0.89       139\n",
            "           5       0.83      0.84      0.84       135\n",
            "           6       0.85      0.89      0.87       130\n",
            "\n",
            "    accuracy                           0.90       924\n",
            "   macro avg       0.90      0.90      0.90       924\n",
            "weighted avg       0.90      0.90      0.90       924\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_6gzlr7B_4o3",
        "colab_type": "text"
      },
      "source": [
        "SVM MODEL"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T7zCeEXhV2hF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "30508ec1-83d8-43fd-85d4-752d12b4c66e"
      },
      "source": [
        "# Building the SVM model \n",
        "svmclf = SVC() \n",
        "svmclf.fit(X_train, y_train) \n",
        "  \n",
        "# Storing the predictions of the non-linear model \n",
        "y_pred_svmclf = svmclf.predict(X_test) \n",
        "  \n",
        "# Evaluating the performance of the non-linear model \n",
        "print('Accuracy : '+str(accuracy_score(y_test, y_pred_svmclf))) "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy : 0.8614718614718615\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iGXEUreF2m_K",
        "colab_type": "text"
      },
      "source": [
        "CNN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_HY_RGHg2llz",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "88028523-d6df-4bc2-d3b4-38c0baf5fde2"
      },
      "source": [
        "x_traincnn = np.expand_dims(X_train, axis=2)\n",
        "x_testcnn = np.expand_dims(X_test, axis=2)\n",
        "x_traincnn.shape, x_testcnn.shape, y_train.shape, y_test.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((1876, 128, 1), (924, 128, 1), (1876,), (924,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "unTT4aPe2jpp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 370
        },
        "outputId": "4f7a6c7f-0b2e-4265-d0f1-039ed2d49712"
      },
      "source": [
        "import keras\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "from keras.preprocessing import sequence\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Embedding\n",
        "from keras.utils import to_categorical\n",
        "from keras.layers import Input, Flatten, Dropout, Activation\n",
        "from keras.layers import Conv1D, MaxPooling1D\n",
        "from keras.models import Model\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Conv1D(32, 5,padding='same',\n",
        "                 input_shape=(128,1)))\n",
        "#model.add(Conv1D(64, 5,padding='same', activation='relu'))\n",
        "#model.add(Activation('relu'))\n",
        "model.add(MaxPooling1D(pool_size=(8)))\n",
        "model.add(Dropout(0.25))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(128, activation='relu'))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(7, activation='softmax'))\n",
        "opt = keras.optimizers.rmsprop(lr=0.00005, rho=0.9, epsilon=None, decay=0.0)\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv1d_25 (Conv1D)           (None, 128, 32)           192       \n",
            "_________________________________________________________________\n",
            "max_pooling1d_15 (MaxPooling (None, 16, 32)            0         \n",
            "_________________________________________________________________\n",
            "dropout_27 (Dropout)         (None, 16, 32)            0         \n",
            "_________________________________________________________________\n",
            "flatten_15 (Flatten)         (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_78 (Dense)             (None, 128)               65664     \n",
            "_________________________________________________________________\n",
            "dropout_28 (Dropout)         (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_79 (Dense)             (None, 7)                 903       \n",
            "=================================================================\n",
            "Total params: 66,759\n",
            "Trainable params: 66,759\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jLt5Iz9y3FTr",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 706
        },
        "outputId": "378f7d29-7331-41c2-b98d-f211c2c7451c"
      },
      "source": [
        "model.compile(loss='sparse_categorical_crossentropy',\n",
        "              optimizer=keras.optimizers.Adadelta(),\n",
        "              metrics=['accuracy'])\n",
        "cnnhistory=model.fit(x_traincnn, y_train, batch_size=128, epochs=20, validation_data=(x_testcnn, y_test))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 1876 samples, validate on 924 samples\n",
            "Epoch 1/20\n",
            "1876/1876 [==============================] - 3s 2ms/step - loss: 11.9277 - acc: 0.1599 - val_loss: 8.0467 - val_acc: 0.2424\n",
            "Epoch 2/20\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 10.4389 - acc: 0.2068 - val_loss: 5.6996 - val_acc: 0.2327\n",
            "Epoch 3/20\n",
            "1876/1876 [==============================] - 0s 131us/step - loss: 6.9491 - acc: 0.2767 - val_loss: 1.6194 - val_acc: 0.4037\n",
            "Epoch 4/20\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 2.7833 - acc: 0.3412 - val_loss: 1.7265 - val_acc: 0.3907\n",
            "Epoch 5/20\n",
            "1876/1876 [==============================] - 0s 127us/step - loss: 1.8117 - acc: 0.3662 - val_loss: 1.5672 - val_acc: 0.5162\n",
            "Epoch 6/20\n",
            "1876/1876 [==============================] - 0s 120us/step - loss: 1.6563 - acc: 0.3886 - val_loss: 1.3054 - val_acc: 0.5931\n",
            "Epoch 7/20\n",
            "1876/1876 [==============================] - 0s 130us/step - loss: 1.5105 - acc: 0.4232 - val_loss: 1.1056 - val_acc: 0.6429\n",
            "Epoch 8/20\n",
            "1876/1876 [==============================] - 0s 120us/step - loss: 1.3640 - acc: 0.4701 - val_loss: 0.8806 - val_acc: 0.7911\n",
            "Epoch 9/20\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 1.2662 - acc: 0.4963 - val_loss: 0.8580 - val_acc: 0.7641\n",
            "Epoch 10/20\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 1.1336 - acc: 0.5384 - val_loss: 0.6015 - val_acc: 0.8409\n",
            "Epoch 11/20\n",
            "1876/1876 [==============================] - 0s 128us/step - loss: 1.0580 - acc: 0.5736 - val_loss: 0.5613 - val_acc: 0.8420\n",
            "Epoch 12/20\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.9351 - acc: 0.6365 - val_loss: 0.4558 - val_acc: 0.9145\n",
            "Epoch 13/20\n",
            "1876/1876 [==============================] - 0s 120us/step - loss: 0.8690 - acc: 0.6546 - val_loss: 0.4244 - val_acc: 0.8972\n",
            "Epoch 14/20\n",
            "1876/1876 [==============================] - 0s 122us/step - loss: 0.8236 - acc: 0.6807 - val_loss: 0.3932 - val_acc: 0.8939\n",
            "Epoch 15/20\n",
            "1876/1876 [==============================] - 0s 122us/step - loss: 0.7416 - acc: 0.6935 - val_loss: 0.3491 - val_acc: 0.9188\n",
            "Epoch 16/20\n",
            "1876/1876 [==============================] - 0s 121us/step - loss: 0.6604 - acc: 0.7377 - val_loss: 0.3211 - val_acc: 0.9188\n",
            "Epoch 17/20\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 0.6551 - acc: 0.7425 - val_loss: 0.3076 - val_acc: 0.9513\n",
            "Epoch 18/20\n",
            "1876/1876 [==============================] - 0s 120us/step - loss: 0.6307 - acc: 0.7516 - val_loss: 0.3292 - val_acc: 0.8961\n",
            "Epoch 19/20\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 0.5742 - acc: 0.7687 - val_loss: 0.2621 - val_acc: 0.9589\n",
            "Epoch 20/20\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.5647 - acc: 0.7729 - val_loss: 0.2521 - val_acc: 0.9372\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tPXlUTHp3sYv",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "outputId": "b620dd93-6c52-4f3a-8885-ea2e3bd0f78b"
      },
      "source": [
        "score = model.evaluate(x_testcnn, y_test, verbose=0)\n",
        "print('Test loss:', score[0])\n",
        "print('Test accuracy:', score[1])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test loss: 0.25205654486433254\n",
            "Test accuracy: 0.9372294377454947\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t3X5YE4a31v9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 722
        },
        "outputId": "134a0c8c-726a-4e06-d668-aedef2771f44"
      },
      "source": [
        "predictions = model.predict_classes(x_testcnn)\n",
        "predictions\n",
        "y_test\n",
        "new_Ytest = y_test.astype(int)\n",
        "new_Ytest"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([3, 0, 3, 3, 2, 1, 4, 0, 4, 3, 5, 5, 3, 3, 3, 6, 0, 6, 5, 3, 4, 0,\n",
              "       6, 4, 0, 6, 5, 1, 0, 4, 1, 6, 0, 0, 3, 6, 5, 1, 0, 6, 2, 0, 3, 5,\n",
              "       6, 6, 0, 4, 4, 0, 1, 1, 1, 6, 3, 3, 2, 4, 2, 1, 3, 4, 6, 4, 5, 4,\n",
              "       0, 1, 4, 3, 3, 6, 1, 5, 0, 6, 4, 6, 0, 2, 0, 2, 0, 4, 6, 1, 1, 5,\n",
              "       6, 6, 5, 0, 3, 4, 0, 6, 1, 3, 3, 3, 3, 4, 5, 5, 4, 0, 4, 2, 0, 6,\n",
              "       1, 1, 4, 2, 0, 6, 0, 2, 3, 6, 3, 6, 4, 0, 3, 4, 2, 4, 0, 5, 3, 6,\n",
              "       5, 5, 5, 2, 3, 2, 3, 1, 2, 3, 4, 4, 2, 4, 4, 2, 1, 3, 6, 3, 4, 4,\n",
              "       5, 5, 0, 0, 3, 1, 1, 6, 4, 5, 1, 1, 1, 0, 2, 5, 0, 1, 1, 6, 1, 3,\n",
              "       4, 2, 5, 2, 1, 3, 2, 3, 4, 6, 6, 1, 4, 5, 6, 0, 5, 1, 6, 5, 4, 4,\n",
              "       6, 5, 3, 3, 6, 6, 0, 6, 2, 6, 3, 4, 2, 0, 3, 4, 6, 2, 2, 0, 4, 4,\n",
              "       6, 5, 2, 1, 3, 4, 0, 0, 3, 6, 4, 4, 2, 3, 2, 5, 6, 2, 3, 5, 4, 1,\n",
              "       0, 2, 1, 0, 1, 1, 4, 4, 3, 3, 2, 2, 3, 5, 3, 0, 6, 1, 3, 4, 5, 0,\n",
              "       0, 6, 5, 4, 3, 0, 3, 1, 4, 6, 0, 6, 2, 6, 1, 4, 4, 0, 4, 2, 2, 5,\n",
              "       2, 2, 6, 4, 5, 5, 3, 4, 4, 1, 4, 4, 2, 1, 4, 0, 3, 2, 4, 2, 3, 0,\n",
              "       1, 4, 5, 4, 3, 6, 3, 1, 5, 1, 4, 3, 2, 3, 2, 6, 4, 5, 6, 3, 2, 4,\n",
              "       3, 5, 6, 3, 5, 4, 2, 6, 0, 5, 5, 2, 0, 6, 5, 5, 6, 2, 1, 2, 4, 4,\n",
              "       4, 5, 6, 0, 6, 2, 1, 3, 6, 5, 4, 1, 0, 4, 3, 3, 5, 1, 0, 0, 5, 2,\n",
              "       2, 5, 1, 1, 4, 2, 1, 3, 4, 2, 3, 0, 6, 5, 4, 6, 2, 1, 1, 3, 1, 6,\n",
              "       2, 2, 0, 5, 5, 4, 3, 1, 1, 0, 2, 5, 0, 0, 6, 4, 5, 3, 6, 5, 4, 6,\n",
              "       2, 2, 4, 3, 6, 6, 6, 4, 6, 5, 5, 5, 4, 3, 2, 2, 0, 2, 0, 2, 1, 6,\n",
              "       0, 6, 6, 1, 0, 0, 5, 6, 5, 5, 0, 4, 5, 2, 2, 4, 6, 0, 5, 3, 6, 6,\n",
              "       3, 3, 0, 5, 6, 3, 3, 1, 1, 3, 4, 5, 6, 3, 5, 6, 5, 4, 4, 6, 2, 1,\n",
              "       0, 4, 0, 2, 1, 5, 1, 6, 5, 4, 5, 4, 2, 6, 5, 2, 5, 5, 4, 0, 1, 3,\n",
              "       4, 1, 4, 4, 5, 6, 3, 6, 6, 3, 0, 5, 4, 0, 1, 2, 4, 6, 0, 3, 1, 5,\n",
              "       3, 3, 0, 6, 2, 3, 6, 1, 0, 2, 1, 6, 1, 5, 5, 2, 2, 2, 2, 5, 1, 2,\n",
              "       1, 3, 5, 3, 3, 0, 3, 3, 0, 4, 6, 1, 2, 1, 0, 1, 5, 6, 1, 3, 2, 5,\n",
              "       0, 1, 3, 3, 5, 5, 4, 2, 1, 2, 0, 5, 3, 3, 6, 6, 6, 4, 6, 1, 1, 4,\n",
              "       1, 2, 5, 1, 2, 4, 5, 1, 2, 5, 0, 6, 3, 3, 5, 0, 5, 6, 3, 3, 4, 1,\n",
              "       4, 4, 2, 1, 5, 1, 0, 5, 5, 0, 1, 1, 0, 4, 2, 3, 3, 6, 5, 0, 3, 4,\n",
              "       1, 2, 0, 3, 2, 0, 6, 2, 1, 0, 5, 3, 0, 0, 6, 4, 3, 1, 5, 3, 0, 1,\n",
              "       0, 4, 0, 6, 6, 2, 5, 6, 6, 2, 2, 0, 6, 3, 1, 0, 2, 4, 3, 0, 0, 3,\n",
              "       1, 4, 6, 1, 1, 5, 0, 6, 5, 4, 1, 2, 2, 0, 6, 0, 2, 2, 5, 0, 2, 3,\n",
              "       0, 3, 0, 0, 5, 1, 4, 2, 1, 3, 5, 0, 2, 5, 3, 4, 3, 3, 1, 4, 5, 3,\n",
              "       3, 5, 3, 6, 6, 6, 4, 0, 5, 5, 3, 3, 3, 3, 1, 1, 6, 0, 4, 6, 3, 4,\n",
              "       1, 4, 0, 0, 1, 6, 1, 4, 4, 1, 5, 1, 0, 1, 2, 3, 2, 3, 1, 0, 6, 5,\n",
              "       1, 3, 6, 2, 5, 3, 4, 5, 6, 2, 6, 3, 2, 1, 5, 6, 0, 6, 4, 2, 4, 2,\n",
              "       0, 5, 3, 5, 1, 4, 3, 6, 6, 6, 6, 0, 4, 6, 0, 2, 2, 2, 5, 5, 6, 4,\n",
              "       5, 5, 6, 1, 4, 2, 0, 0, 0, 2, 3, 5, 4, 2, 1, 3, 2, 4, 4, 0, 2, 0,\n",
              "       1, 6, 5, 2, 5, 0, 1, 0, 5, 2, 2, 4, 1, 4, 3, 4, 5, 1, 5, 4, 1, 0,\n",
              "       5, 4, 6, 4, 1, 4, 1, 4, 2, 1, 2, 2, 4, 2, 4, 5, 0, 1, 2, 5, 2, 3,\n",
              "       2, 0, 5, 1, 0, 5, 1, 5, 5, 3, 2, 6, 6, 2, 2, 5, 2, 5, 4, 4, 3, 1,\n",
              "       1, 4, 3, 5, 4, 6, 0, 0, 0, 5, 1, 2, 0, 3, 2, 6, 2, 3, 2, 5, 6, 3])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 136
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RXodqiDa4BYU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 252
        },
        "outputId": "405cafb5-ba59-4f7f-93f7-b53890510699"
      },
      "source": [
        "from sklearn.metrics import classification_report\n",
        "report = classification_report(new_Ytest, predictions)\n",
        "print(report)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      1.00      1.00       128\n",
            "           1       1.00      0.98      0.99       125\n",
            "           2       0.88      0.97      0.92       130\n",
            "           3       0.96      1.00      0.98       137\n",
            "           4       1.00      0.96      0.98       139\n",
            "           5       0.96      0.79      0.87       135\n",
            "           6       0.78      0.87      0.82       130\n",
            "\n",
            "    accuracy                           0.94       924\n",
            "   macro avg       0.94      0.94      0.94       924\n",
            "weighted avg       0.94      0.94      0.94       924\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "veALfo4d_YYm",
        "colab_type": "text"
      },
      "source": [
        "Confusion matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mkiqLKSi_MF3",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 134
        },
        "outputId": "7de1e9c9-3ec4-40ac-e9bb-8eb2ffe0b848"
      },
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "matrix = confusion_matrix(new_Ytest, predictions)\n",
        "print (matrix)\n",
        "\n",
        "# 0 = angry, 1 = disgust, 2 = fear, 3 = happy, 4 = neutral, 5 = surprise, 6 = sad"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[128   0   0   0   0   0   0]\n",
            " [  0 122   0   0   0   3   0]\n",
            " [  0   0 126   1   0   1   2]\n",
            " [  0   0   0 137   0   0   0]\n",
            " [  0   0   0   5 134   0   0]\n",
            " [  0   0   0   0   0 106  29]\n",
            " [  0   0  17   0   0   0 113]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JqRtqmYN_07r",
        "colab_type": "text"
      },
      "source": [
        "Save the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3luRSyeA_akh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "88cc31b7-1f13-469d-c0e1-687929004693"
      },
      "source": [
        "model_name = 'Emotion_Voice_Detection_Model_TORONTO_ORIGINAL.h5'\n",
        "save_dir = '/content/drive/My Drive/TORONTO_model'\n",
        "# Save model and weights\n",
        "if not os.path.isdir(save_dir):\n",
        "    os.makedirs(save_dir)\n",
        "model_path = os.path.join(save_dir, model_name)\n",
        "model.save(model_path)\n",
        "print('Saved trained model at %s ' % model_path)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Saved trained model at /content/drive/My Drive/TORONTO_model/Emotion_Voice_Detection_Model_TORONTO_ORIGINAL.h5 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gFpOuacq_lC8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 370
        },
        "outputId": "a3b9152f-8519-48d0-e3a4-2482c6d21a7d"
      },
      "source": [
        "loaded_model = keras.models.load_model('/content/drive/My Drive/TORONTO_model/Emotion_Voice_Detection_Model_TORONTO_ORIGINAL.h5')\n",
        "loaded_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv1d_25 (Conv1D)           (None, 128, 32)           192       \n",
            "_________________________________________________________________\n",
            "max_pooling1d_15 (MaxPooling (None, 16, 32)            0         \n",
            "_________________________________________________________________\n",
            "dropout_27 (Dropout)         (None, 16, 32)            0         \n",
            "_________________________________________________________________\n",
            "flatten_15 (Flatten)         (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_78 (Dense)             (None, 128)               65664     \n",
            "_________________________________________________________________\n",
            "dropout_28 (Dropout)         (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_79 (Dense)             (None, 7)                 903       \n",
            "=================================================================\n",
            "Total params: 66,759\n",
            "Trainable params: 66,759\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0Xivmni8_utx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "outputId": "c18aa088-2278-4cdc-8f7c-db74a16b6c52"
      },
      "source": [
        "loss, acc = loaded_model.evaluate(x_testcnn, y_test)\n",
        "print(\"Restored model, accuracy: {:5.2f}%\".format(100*acc))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "924/924 [==============================] - 1s 1ms/step\n",
            "Restored model, accuracy: 93.72%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4T-vZ49yX4Ao",
        "colab_type": "text"
      },
      "source": [
        "APPLYING AUTOENCODER"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y-bCRZqEYQTQ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 558
        },
        "outputId": "89b6f889-a8f3-4850-b01d-404c3a6e63e6"
      },
      "source": [
        "!pip install pydub\n",
        "from scipy.io import wavfile\n",
        "import wave\n",
        "import struct\n",
        "from struct import *\n",
        "import matplotlib.pyplot as plt\n",
        "!pip install keras==2.1.5\n",
        "from keras.layers import Input, Dense\n",
        "from keras.models import Model\n",
        "import matplotlib.pyplot as plt\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "from sklearn.preprocessing import MinMaxScaler"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: pydub in /usr/local/lib/python3.6/dist-packages (0.23.1)\n",
            "Requirement already satisfied: keras==2.1.5 in /usr/local/lib/python3.6/dist-packages (2.1.5)\n",
            "Requirement already satisfied: pyyaml in /usr/local/lib/python3.6/dist-packages (from keras==2.1.5) (3.13)\n",
            "Requirement already satisfied: six>=1.9.0 in /usr/local/lib/python3.6/dist-packages (from keras==2.1.5) (1.12.0)\n",
            "Requirement already satisfied: numpy>=1.9.1 in /usr/local/lib/python3.6/dist-packages (from keras==2.1.5) (1.18.3)\n",
            "Requirement already satisfied: scipy>=0.14 in /usr/local/lib/python3.6/dist-packages (from keras==2.1.5) (1.4.1)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n",
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
            "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i5VW9XVCWDU4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "x_train= MinMaxScaler().fit_transform(lst)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oJEOykS0YaC4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "fc709fec-23b6-42ad-9098-e12cf148a0ca"
      },
      "source": [
        "x_train.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2800, 128)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Dzmz-fnHaKq8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "a9531fda-c32e-4d9d-9fff-cadbb9cd055f"
      },
      "source": [
        "encoding_dim = 64\n",
        "input_img = Input(shape=(128,))\n",
        "encoded = Dense(encoding_dim, activation='relu')(input_img)\n",
        "\n",
        "\n",
        "decoded = Dense(128, activation='sigmoid')(encoded)\n",
        "\n",
        "autoencoder = Model(input_img, decoded)\n",
        "encoder = Model(input=input_img, output=encoded)\n",
        "encoded_input = Input(shape=(encoding_dim,))\n",
        "\n",
        "decoder_layer = autoencoder.layers[-1]\n",
        "\n",
        "decoder = Model(input=encoded_input, output=decoder_layer(encoded_input))\n",
        "autoencoder.compile(optimizer='adadelta', loss='binary_crossentropy')\n",
        "\n",
        "\n",
        "filepath=\"audio.compress-{loss:.4f}.hdf5\"\n",
        "checkpoint = ModelCheckpoint(filepath, monitor='binary_crossentropy', verbose=1, save_best_only=False)\n",
        "autoencoder.summary()\n",
        "autoencoder.fit(x_train, x_train,\n",
        "                epochs=100,\n",
        "                batch_size=256,\n",
        "                shuffle=True,\n",
        "                validation_data=(x_train, x_train),callbacks=[checkpoint],verbose=0)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:13: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"de...)`\n",
            "  del sys.path[0]\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:18: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"de...)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_19 (InputLayer)        (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_52 (Dense)             (None, 64)                8256      \n",
            "_________________________________________________________________\n",
            "dense_53 (Dense)             (None, 128)               8320      \n",
            "=================================================================\n",
            "Total params: 16,576\n",
            "Trainable params: 16,576\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "\n",
            "Epoch 00001: saving model to audio.compress-0.6981.hdf5\n",
            "\n",
            "Epoch 00002: saving model to audio.compress-0.6894.hdf5\n",
            "\n",
            "Epoch 00003: saving model to audio.compress-0.6851.hdf5\n",
            "\n",
            "Epoch 00004: saving model to audio.compress-0.6813.hdf5\n",
            "\n",
            "Epoch 00005: saving model to audio.compress-0.6777.hdf5\n",
            "\n",
            "Epoch 00006: saving model to audio.compress-0.6747.hdf5\n",
            "\n",
            "Epoch 00007: saving model to audio.compress-0.6724.hdf5\n",
            "\n",
            "Epoch 00008: saving model to audio.compress-0.6709.hdf5\n",
            "\n",
            "Epoch 00009: saving model to audio.compress-0.6699.hdf5\n",
            "\n",
            "Epoch 00010: saving model to audio.compress-0.6693.hdf5\n",
            "\n",
            "Epoch 00011: saving model to audio.compress-0.6689.hdf5\n",
            "\n",
            "Epoch 00012: saving model to audio.compress-0.6687.hdf5\n",
            "\n",
            "Epoch 00013: saving model to audio.compress-0.6685.hdf5\n",
            "\n",
            "Epoch 00014: saving model to audio.compress-0.6684.hdf5\n",
            "\n",
            "Epoch 00015: saving model to audio.compress-0.6683.hdf5\n",
            "\n",
            "Epoch 00016: saving model to audio.compress-0.6682.hdf5\n",
            "\n",
            "Epoch 00017: saving model to audio.compress-0.6681.hdf5\n",
            "\n",
            "Epoch 00018: saving model to audio.compress-0.6680.hdf5\n",
            "\n",
            "Epoch 00019: saving model to audio.compress-0.6679.hdf5\n",
            "\n",
            "Epoch 00020: saving model to audio.compress-0.6678.hdf5\n",
            "\n",
            "Epoch 00021: saving model to audio.compress-0.6677.hdf5\n",
            "\n",
            "Epoch 00022: saving model to audio.compress-0.6677.hdf5\n",
            "\n",
            "Epoch 00023: saving model to audio.compress-0.6676.hdf5\n",
            "\n",
            "Epoch 00024: saving model to audio.compress-0.6675.hdf5\n",
            "\n",
            "Epoch 00025: saving model to audio.compress-0.6674.hdf5\n",
            "\n",
            "Epoch 00026: saving model to audio.compress-0.6673.hdf5\n",
            "\n",
            "Epoch 00027: saving model to audio.compress-0.6672.hdf5\n",
            "\n",
            "Epoch 00028: saving model to audio.compress-0.6671.hdf5\n",
            "\n",
            "Epoch 00029: saving model to audio.compress-0.6670.hdf5\n",
            "\n",
            "Epoch 00030: saving model to audio.compress-0.6669.hdf5\n",
            "\n",
            "Epoch 00031: saving model to audio.compress-0.6667.hdf5\n",
            "\n",
            "Epoch 00032: saving model to audio.compress-0.6666.hdf5\n",
            "\n",
            "Epoch 00033: saving model to audio.compress-0.6665.hdf5\n",
            "\n",
            "Epoch 00034: saving model to audio.compress-0.6664.hdf5\n",
            "\n",
            "Epoch 00035: saving model to audio.compress-0.6663.hdf5\n",
            "\n",
            "Epoch 00036: saving model to audio.compress-0.6662.hdf5\n",
            "\n",
            "Epoch 00037: saving model to audio.compress-0.6660.hdf5\n",
            "\n",
            "Epoch 00038: saving model to audio.compress-0.6659.hdf5\n",
            "\n",
            "Epoch 00039: saving model to audio.compress-0.6658.hdf5\n",
            "\n",
            "Epoch 00040: saving model to audio.compress-0.6657.hdf5\n",
            "\n",
            "Epoch 00041: saving model to audio.compress-0.6655.hdf5\n",
            "\n",
            "Epoch 00042: saving model to audio.compress-0.6654.hdf5\n",
            "\n",
            "Epoch 00043: saving model to audio.compress-0.6653.hdf5\n",
            "\n",
            "Epoch 00044: saving model to audio.compress-0.6651.hdf5\n",
            "\n",
            "Epoch 00045: saving model to audio.compress-0.6650.hdf5\n",
            "\n",
            "Epoch 00046: saving model to audio.compress-0.6648.hdf5\n",
            "\n",
            "Epoch 00047: saving model to audio.compress-0.6647.hdf5\n",
            "\n",
            "Epoch 00048: saving model to audio.compress-0.6645.hdf5\n",
            "\n",
            "Epoch 00049: saving model to audio.compress-0.6644.hdf5\n",
            "\n",
            "Epoch 00050: saving model to audio.compress-0.6642.hdf5\n",
            "\n",
            "Epoch 00051: saving model to audio.compress-0.6641.hdf5\n",
            "\n",
            "Epoch 00052: saving model to audio.compress-0.6639.hdf5\n",
            "\n",
            "Epoch 00053: saving model to audio.compress-0.6637.hdf5\n",
            "\n",
            "Epoch 00054: saving model to audio.compress-0.6636.hdf5\n",
            "\n",
            "Epoch 00055: saving model to audio.compress-0.6634.hdf5\n",
            "\n",
            "Epoch 00056: saving model to audio.compress-0.6632.hdf5\n",
            "\n",
            "Epoch 00057: saving model to audio.compress-0.6631.hdf5\n",
            "\n",
            "Epoch 00058: saving model to audio.compress-0.6629.hdf5\n",
            "\n",
            "Epoch 00059: saving model to audio.compress-0.6627.hdf5\n",
            "\n",
            "Epoch 00060: saving model to audio.compress-0.6625.hdf5\n",
            "\n",
            "Epoch 00061: saving model to audio.compress-0.6624.hdf5\n",
            "\n",
            "Epoch 00062: saving model to audio.compress-0.6622.hdf5\n",
            "\n",
            "Epoch 00063: saving model to audio.compress-0.6620.hdf5\n",
            "\n",
            "Epoch 00064: saving model to audio.compress-0.6618.hdf5\n",
            "\n",
            "Epoch 00065: saving model to audio.compress-0.6616.hdf5\n",
            "\n",
            "Epoch 00066: saving model to audio.compress-0.6615.hdf5\n",
            "\n",
            "Epoch 00067: saving model to audio.compress-0.6613.hdf5\n",
            "\n",
            "Epoch 00068: saving model to audio.compress-0.6611.hdf5\n",
            "\n",
            "Epoch 00069: saving model to audio.compress-0.6609.hdf5\n",
            "\n",
            "Epoch 00070: saving model to audio.compress-0.6607.hdf5\n",
            "\n",
            "Epoch 00071: saving model to audio.compress-0.6606.hdf5\n",
            "\n",
            "Epoch 00072: saving model to audio.compress-0.6604.hdf5\n",
            "\n",
            "Epoch 00073: saving model to audio.compress-0.6602.hdf5\n",
            "\n",
            "Epoch 00074: saving model to audio.compress-0.6600.hdf5\n",
            "\n",
            "Epoch 00075: saving model to audio.compress-0.6599.hdf5\n",
            "\n",
            "Epoch 00076: saving model to audio.compress-0.6597.hdf5\n",
            "\n",
            "Epoch 00077: saving model to audio.compress-0.6595.hdf5\n",
            "\n",
            "Epoch 00078: saving model to audio.compress-0.6593.hdf5\n",
            "\n",
            "Epoch 00079: saving model to audio.compress-0.6592.hdf5\n",
            "\n",
            "Epoch 00080: saving model to audio.compress-0.6590.hdf5\n",
            "\n",
            "Epoch 00081: saving model to audio.compress-0.6588.hdf5\n",
            "\n",
            "Epoch 00082: saving model to audio.compress-0.6587.hdf5\n",
            "\n",
            "Epoch 00083: saving model to audio.compress-0.6585.hdf5\n",
            "\n",
            "Epoch 00084: saving model to audio.compress-0.6583.hdf5\n",
            "\n",
            "Epoch 00085: saving model to audio.compress-0.6582.hdf5\n",
            "\n",
            "Epoch 00086: saving model to audio.compress-0.6580.hdf5\n",
            "\n",
            "Epoch 00087: saving model to audio.compress-0.6579.hdf5\n",
            "\n",
            "Epoch 00088: saving model to audio.compress-0.6577.hdf5\n",
            "\n",
            "Epoch 00089: saving model to audio.compress-0.6576.hdf5\n",
            "\n",
            "Epoch 00090: saving model to audio.compress-0.6574.hdf5\n",
            "\n",
            "Epoch 00091: saving model to audio.compress-0.6573.hdf5\n",
            "\n",
            "Epoch 00092: saving model to audio.compress-0.6571.hdf5\n",
            "\n",
            "Epoch 00093: saving model to audio.compress-0.6570.hdf5\n",
            "\n",
            "Epoch 00094: saving model to audio.compress-0.6569.hdf5\n",
            "\n",
            "Epoch 00095: saving model to audio.compress-0.6567.hdf5\n",
            "\n",
            "Epoch 00096: saving model to audio.compress-0.6566.hdf5\n",
            "\n",
            "Epoch 00097: saving model to audio.compress-0.6565.hdf5\n",
            "\n",
            "Epoch 00098: saving model to audio.compress-0.6563.hdf5\n",
            "\n",
            "Epoch 00099: saving model to audio.compress-0.6562.hdf5\n",
            "\n",
            "Epoch 00100: saving model to audio.compress-0.6561.hdf5\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f6cf2a9db70>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D4QZ-RFGrX7x",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 578
        },
        "outputId": "4e2d1f6c-fe3d-4b09-ec9d-9dbc00a5e935"
      },
      "source": [
        "filename = \"audio.compress-0.6561.hdf5\"\n",
        "autoencoder.load_weights(filename)\n",
        "autoencoder.compile(loss='mean_squared_error', optimizer='adam')\n",
        "encoded_imgs = encoder.predict(x_train)\n",
        "decoded_imgs = decoder.predict(encoded_imgs)\n",
        "\n",
        "n = 1  # how many files we will display\n",
        "plt.figure(figsize=(10, 5))\n",
        "for i in range(n):\n",
        "    # display original\n",
        "    ax = plt.subplot(2, n, i + 1)\n",
        "    plt.imshow(x_train[i].reshape(16,8 ))\n",
        "    plt.title(\"ORIGINAL AUDIO\")\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "plt.show()\n",
        "\n",
        "plt.figure(figsize=(8, .5))\n",
        "for i in range(n):\n",
        "    ax = plt.subplot(1, n, i+1)\n",
        "    plt.imshow(encoded_imgs.reshape(1792,100).T)\n",
        "    plt.title('ENCODED AUDIO')    \n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "plt.show()\n",
        "\n",
        "plt.figure(figsize=(10, 5))\n",
        "for i in range(n):\n",
        "    # display reconstruction\n",
        "    ax = plt.subplot(2, n, i + 1 + n)\n",
        "    plt.imshow(decoded_imgs[i].reshape(16,8 ))\n",
        "    plt.title(\"RECONSTRUCTED AUDIO\")\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "plt.show()\n",
        "\n",
        "print(\"Accuracy=\",1-np.mean(abs(x_train-decoded_imgs)),'\\n')\n",
        "\n",
        "plt.figure(figsize=(10, 2))\n",
        "plt.plot(x_train[0],color='r',linewidth=3)\n",
        "plt.plot(decoded_imgs[0])\n",
        "plt.title('ORIGINAL (blue) and RECONSTRUCTED AUDIO (red)')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x36 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Accuracy= 0.9192111867933958 \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x144 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tKOA1fhTuy9d",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "cd8a9797-e5d2-4dc4-9f66-6e9dd2bddf28"
      },
      "source": [
        "decoded_imgs.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2800, 128)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q4RaVDNnu3HD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train_encoded, X_test_encoded, y_train_encoded, y_test_encoded = train_test_split(decoded_imgs, y, test_size=0.33, random_state=42)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tW3T0Ncr-tpy",
        "colab_type": "text"
      },
      "source": [
        "SVM MODEL"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wlUX-Es8vdZA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "9789de7d-dcab-4093-ddc4-6f64364999ae"
      },
      "source": [
        "# Building the SVM model \n",
        "svmclf = SVC() \n",
        "svmclf.fit(X_train_encoded, y_train_encoded) \n",
        "  \n",
        "# Storing the predictions of the non-linear model \n",
        "y_pred_svmclf = svmclf.predict(X_test_encoded) \n",
        "  \n",
        "# Evaluating the performance of the non-linear model \n",
        "print('Accuracy : '+str(accuracy_score(y_test_encoded, y_pred_svmclf))) "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy : 0.9199134199134199\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LW8dnHsf-pnp",
        "colab_type": "text"
      },
      "source": [
        "DECISION TREE CLASSIFIER"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U035_Yigv3Pf",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 252
        },
        "outputId": "a8e56269-b5a6-409a-b2ba-e3d611a4e701"
      },
      "source": [
        "from sklearn.tree import DecisionTreeClassifier\n",
        "\n",
        "dtree = DecisionTreeClassifier()\n",
        "\n",
        "dtree.fit(X_train_encoded, y_train_encoded)\n",
        "\n",
        "predictions = dtree.predict(X_test_encoded)\n",
        "from sklearn.metrics import classification_report,confusion_matrix\n",
        "print(classification_report(y_test_encoded,predictions))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.93      0.97      0.95       128\n",
            "           1       0.94      0.97      0.95       125\n",
            "           2       0.89      0.87      0.88       130\n",
            "           3       0.90      0.86      0.88       137\n",
            "           4       0.86      0.90      0.88       139\n",
            "           5       0.87      0.84      0.86       135\n",
            "           6       0.89      0.87      0.88       130\n",
            "\n",
            "    accuracy                           0.90       924\n",
            "   macro avg       0.90      0.90      0.90       924\n",
            "weighted avg       0.90      0.90      0.90       924\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BcYosM__wwkz",
        "colab_type": "text"
      },
      "source": [
        "CNN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "clRq2cylwuuu",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "5e080817-2c61-4bf0-9e71-d0db946da5fa"
      },
      "source": [
        "x_traincnn_encoded = np.expand_dims(X_train_encoded, axis=2)\n",
        "x_testcnn_encoded = np.expand_dims(X_test_encoded, axis=2)\n",
        "x_traincnn_encoded.shape, x_testcnn_encoded.shape, y_train_encoded.shape, y_test_encoded.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((1876, 128, 1), (924, 128, 1), (1876,), (924,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bQ-i5zF6xQ_j",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 370
        },
        "outputId": "cd85529e-2c2e-4e15-d88b-1ac6874e769b"
      },
      "source": [
        "import keras\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "from keras.preprocessing import sequence\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Embedding\n",
        "from keras.utils import to_categorical\n",
        "from keras.layers import Input, Flatten, Dropout, Activation\n",
        "from keras.layers import Conv1D, MaxPooling1D\n",
        "from keras.models import Model\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Conv1D(32, 5,padding='same',\n",
        "                 input_shape=(128,1)))\n",
        "#model.add(Conv1D(64, 5,padding='same', activation='relu'))\n",
        "#model.add(Activation('relu'))\n",
        "model.add(MaxPooling1D(pool_size=(8)))\n",
        "model.add(Dropout(0.25))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(128, activation='relu'))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(7, activation='softmax'))\n",
        "#opt = keras.optimizers.rmsprop(lr=0.00005, rho=0.9, epsilon=None, decay=0.0)\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv1d_24 (Conv1D)           (None, 128, 32)           192       \n",
            "_________________________________________________________________\n",
            "max_pooling1d_14 (MaxPooling (None, 16, 32)            0         \n",
            "_________________________________________________________________\n",
            "dropout_25 (Dropout)         (None, 16, 32)            0         \n",
            "_________________________________________________________________\n",
            "flatten_14 (Flatten)         (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_76 (Dense)             (None, 128)               65664     \n",
            "_________________________________________________________________\n",
            "dropout_26 (Dropout)         (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_77 (Dense)             (None, 7)                 903       \n",
            "=================================================================\n",
            "Total params: 66,759\n",
            "Trainable params: 66,759\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PoUFzN5Jyes5",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "9c979fda-c987-43ef-9647-ef29c94286e8"
      },
      "source": [
        "model.compile(loss='sparse_categorical_crossentropy',\n",
        "              optimizer=keras.optimizers.Adadelta(),\n",
        "              metrics=['accuracy'])\n",
        "cnnhistory_encoded=model.fit(x_traincnn_encoded, y_train_encoded, batch_size=128, epochs=70, validation_data=(x_testcnn_encoded, y_test_encoded))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 1876 samples, validate on 924 samples\n",
            "Epoch 1/70\n",
            "1876/1876 [==============================] - 3s 2ms/step - loss: 0.2103 - acc: 0.9216 - val_loss: 0.1479 - val_acc: 0.9513\n",
            "Epoch 2/70\n",
            "1876/1876 [==============================] - 0s 122us/step - loss: 0.2106 - acc: 0.9254 - val_loss: 0.1466 - val_acc: 0.9535\n",
            "Epoch 3/70\n",
            "1876/1876 [==============================] - 0s 127us/step - loss: 0.1842 - acc: 0.9350 - val_loss: 0.1518 - val_acc: 0.9459\n",
            "Epoch 4/70\n",
            "1876/1876 [==============================] - 0s 128us/step - loss: 0.1995 - acc: 0.9216 - val_loss: 0.1433 - val_acc: 0.9524\n",
            "Epoch 5/70\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 0.2029 - acc: 0.9286 - val_loss: 0.1519 - val_acc: 0.9459\n",
            "Epoch 6/70\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 0.1994 - acc: 0.9296 - val_loss: 0.1480 - val_acc: 0.9491\n",
            "Epoch 7/70\n",
            "1876/1876 [==============================] - 0s 122us/step - loss: 0.2023 - acc: 0.9323 - val_loss: 0.1457 - val_acc: 0.9524\n",
            "Epoch 8/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1888 - acc: 0.9350 - val_loss: 0.1511 - val_acc: 0.9437\n",
            "Epoch 9/70\n",
            "1876/1876 [==============================] - 0s 128us/step - loss: 0.2106 - acc: 0.9211 - val_loss: 0.1488 - val_acc: 0.9437\n",
            "Epoch 10/70\n",
            "1876/1876 [==============================] - 0s 117us/step - loss: 0.1977 - acc: 0.9286 - val_loss: 0.1424 - val_acc: 0.9513\n",
            "Epoch 11/70\n",
            "1876/1876 [==============================] - 0s 117us/step - loss: 0.1856 - acc: 0.9350 - val_loss: 0.1471 - val_acc: 0.9481\n",
            "Epoch 12/70\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 0.2043 - acc: 0.9211 - val_loss: 0.1563 - val_acc: 0.9502\n",
            "Epoch 13/70\n",
            "1876/1876 [==============================] - 0s 131us/step - loss: 0.1979 - acc: 0.9259 - val_loss: 0.1438 - val_acc: 0.9535\n",
            "Epoch 14/70\n",
            "1876/1876 [==============================] - 0s 131us/step - loss: 0.1932 - acc: 0.9264 - val_loss: 0.1562 - val_acc: 0.9426\n",
            "Epoch 15/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1880 - acc: 0.9312 - val_loss: 0.1411 - val_acc: 0.9524\n",
            "Epoch 16/70\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 0.1910 - acc: 0.9312 - val_loss: 0.1544 - val_acc: 0.9481\n",
            "Epoch 17/70\n",
            "1876/1876 [==============================] - 0s 128us/step - loss: 0.1801 - acc: 0.9350 - val_loss: 0.1438 - val_acc: 0.9535\n",
            "Epoch 18/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1840 - acc: 0.9307 - val_loss: 0.1319 - val_acc: 0.9578\n",
            "Epoch 19/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1875 - acc: 0.9334 - val_loss: 0.1390 - val_acc: 0.9524\n",
            "Epoch 20/70\n",
            "1876/1876 [==============================] - 0s 123us/step - loss: 0.1887 - acc: 0.9339 - val_loss: 0.1352 - val_acc: 0.9578\n",
            "Epoch 21/70\n",
            "1876/1876 [==============================] - 0s 127us/step - loss: 0.1822 - acc: 0.9291 - val_loss: 0.1450 - val_acc: 0.9513\n",
            "Epoch 22/70\n",
            "1876/1876 [==============================] - 0s 118us/step - loss: 0.2025 - acc: 0.9259 - val_loss: 0.1362 - val_acc: 0.9578\n",
            "Epoch 23/70\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 0.1758 - acc: 0.9398 - val_loss: 0.1371 - val_acc: 0.9535\n",
            "Epoch 24/70\n",
            "1876/1876 [==============================] - 0s 127us/step - loss: 0.1762 - acc: 0.9398 - val_loss: 0.1371 - val_acc: 0.9502\n",
            "Epoch 25/70\n",
            "1876/1876 [==============================] - 0s 129us/step - loss: 0.1852 - acc: 0.9350 - val_loss: 0.1385 - val_acc: 0.9567\n",
            "Epoch 26/70\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 0.1662 - acc: 0.9424 - val_loss: 0.1289 - val_acc: 0.9621\n",
            "Epoch 27/70\n",
            "1876/1876 [==============================] - 0s 120us/step - loss: 0.1859 - acc: 0.9296 - val_loss: 0.1350 - val_acc: 0.9643\n",
            "Epoch 28/70\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 0.1824 - acc: 0.9371 - val_loss: 0.1353 - val_acc: 0.9513\n",
            "Epoch 29/70\n",
            "1876/1876 [==============================] - 0s 119us/step - loss: 0.1787 - acc: 0.9350 - val_loss: 0.1347 - val_acc: 0.9545\n",
            "Epoch 30/70\n",
            "1876/1876 [==============================] - 0s 131us/step - loss: 0.1781 - acc: 0.9366 - val_loss: 0.1401 - val_acc: 0.9481\n",
            "Epoch 31/70\n",
            "1876/1876 [==============================] - 0s 130us/step - loss: 0.1914 - acc: 0.9248 - val_loss: 0.1297 - val_acc: 0.9524\n",
            "Epoch 32/70\n",
            "1876/1876 [==============================] - 0s 123us/step - loss: 0.1945 - acc: 0.9259 - val_loss: 0.1490 - val_acc: 0.9513\n",
            "Epoch 33/70\n",
            "1876/1876 [==============================] - 0s 132us/step - loss: 0.1758 - acc: 0.9376 - val_loss: 0.1443 - val_acc: 0.9567\n",
            "Epoch 34/70\n",
            "1876/1876 [==============================] - 0s 128us/step - loss: 0.1832 - acc: 0.9350 - val_loss: 0.1327 - val_acc: 0.9535\n",
            "Epoch 35/70\n",
            "1876/1876 [==============================] - 0s 121us/step - loss: 0.1872 - acc: 0.9318 - val_loss: 0.1423 - val_acc: 0.9545\n",
            "Epoch 36/70\n",
            "1876/1876 [==============================] - 0s 122us/step - loss: 0.1742 - acc: 0.9376 - val_loss: 0.1478 - val_acc: 0.9448\n",
            "Epoch 37/70\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 0.1829 - acc: 0.9302 - val_loss: 0.1351 - val_acc: 0.9535\n",
            "Epoch 38/70\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 0.1869 - acc: 0.9254 - val_loss: 0.1420 - val_acc: 0.9491\n",
            "Epoch 39/70\n",
            "1876/1876 [==============================] - 0s 121us/step - loss: 0.1740 - acc: 0.9371 - val_loss: 0.1277 - val_acc: 0.9578\n",
            "Epoch 40/70\n",
            "1876/1876 [==============================] - 0s 119us/step - loss: 0.1885 - acc: 0.9318 - val_loss: 0.1403 - val_acc: 0.9535\n",
            "Epoch 41/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1807 - acc: 0.9392 - val_loss: 0.1473 - val_acc: 0.9405\n",
            "Epoch 42/70\n",
            "1876/1876 [==============================] - 0s 123us/step - loss: 0.1801 - acc: 0.9392 - val_loss: 0.1299 - val_acc: 0.9578\n",
            "Epoch 43/70\n",
            "1876/1876 [==============================] - 0s 132us/step - loss: 0.1786 - acc: 0.9334 - val_loss: 0.1291 - val_acc: 0.9600\n",
            "Epoch 44/70\n",
            "1876/1876 [==============================] - 0s 133us/step - loss: 0.1810 - acc: 0.9296 - val_loss: 0.1401 - val_acc: 0.9524\n",
            "Epoch 45/70\n",
            "1876/1876 [==============================] - 0s 133us/step - loss: 0.1592 - acc: 0.9472 - val_loss: 0.1335 - val_acc: 0.9600\n",
            "Epoch 46/70\n",
            "1876/1876 [==============================] - 0s 140us/step - loss: 0.1655 - acc: 0.9371 - val_loss: 0.1337 - val_acc: 0.9491\n",
            "Epoch 47/70\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 0.1668 - acc: 0.9414 - val_loss: 0.1342 - val_acc: 0.9567\n",
            "Epoch 48/70\n",
            "1876/1876 [==============================] - 0s 120us/step - loss: 0.1709 - acc: 0.9350 - val_loss: 0.1256 - val_acc: 0.9621\n",
            "Epoch 49/70\n",
            "1876/1876 [==============================] - 0s 129us/step - loss: 0.1679 - acc: 0.9414 - val_loss: 0.1422 - val_acc: 0.9545\n",
            "Epoch 50/70\n",
            "1876/1876 [==============================] - 0s 123us/step - loss: 0.1682 - acc: 0.9382 - val_loss: 0.1309 - val_acc: 0.9556\n",
            "Epoch 51/70\n",
            "1876/1876 [==============================] - 0s 130us/step - loss: 0.1821 - acc: 0.9339 - val_loss: 0.1274 - val_acc: 0.9610\n",
            "Epoch 52/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1686 - acc: 0.9382 - val_loss: 0.1409 - val_acc: 0.9524\n",
            "Epoch 53/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1651 - acc: 0.9483 - val_loss: 0.1219 - val_acc: 0.9665\n",
            "Epoch 54/70\n",
            "1876/1876 [==============================] - 0s 131us/step - loss: 0.1746 - acc: 0.9387 - val_loss: 0.1226 - val_acc: 0.9600\n",
            "Epoch 55/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1713 - acc: 0.9451 - val_loss: 0.1251 - val_acc: 0.9643\n",
            "Epoch 56/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1674 - acc: 0.9403 - val_loss: 0.1390 - val_acc: 0.9535\n",
            "Epoch 57/70\n",
            "1876/1876 [==============================] - 0s 122us/step - loss: 0.1826 - acc: 0.9302 - val_loss: 0.1204 - val_acc: 0.9621\n",
            "Epoch 58/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1665 - acc: 0.9430 - val_loss: 0.1287 - val_acc: 0.9589\n",
            "Epoch 59/70\n",
            "1876/1876 [==============================] - 0s 130us/step - loss: 0.1720 - acc: 0.9270 - val_loss: 0.1263 - val_acc: 0.9643\n",
            "Epoch 60/70\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 0.1768 - acc: 0.9403 - val_loss: 0.1299 - val_acc: 0.9545\n",
            "Epoch 61/70\n",
            "1876/1876 [==============================] - 0s 128us/step - loss: 0.1620 - acc: 0.9376 - val_loss: 0.1253 - val_acc: 0.9621\n",
            "Epoch 62/70\n",
            "1876/1876 [==============================] - 0s 132us/step - loss: 0.1586 - acc: 0.9440 - val_loss: 0.1249 - val_acc: 0.9600\n",
            "Epoch 63/70\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 0.1628 - acc: 0.9440 - val_loss: 0.1371 - val_acc: 0.9470\n",
            "Epoch 64/70\n",
            "1876/1876 [==============================] - 0s 127us/step - loss: 0.1710 - acc: 0.9419 - val_loss: 0.1223 - val_acc: 0.9610\n",
            "Epoch 65/70\n",
            "1876/1876 [==============================] - 0s 125us/step - loss: 0.1642 - acc: 0.9456 - val_loss: 0.1169 - val_acc: 0.9654\n",
            "Epoch 66/70\n",
            "1876/1876 [==============================] - 0s 123us/step - loss: 0.1604 - acc: 0.9446 - val_loss: 0.1222 - val_acc: 0.9610\n",
            "Epoch 67/70\n",
            "1876/1876 [==============================] - 0s 124us/step - loss: 0.1680 - acc: 0.9371 - val_loss: 0.1264 - val_acc: 0.9600\n",
            "Epoch 68/70\n",
            "1876/1876 [==============================] - 0s 128us/step - loss: 0.1592 - acc: 0.9392 - val_loss: 0.1214 - val_acc: 0.9600\n",
            "Epoch 69/70\n",
            "1876/1876 [==============================] - 0s 121us/step - loss: 0.1583 - acc: 0.9483 - val_loss: 0.1243 - val_acc: 0.9643\n",
            "Epoch 70/70\n",
            "1876/1876 [==============================] - 0s 126us/step - loss: 0.1653 - acc: 0.9446 - val_loss: 0.1119 - val_acc: 0.9610\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M7QaW9oR157S",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "outputId": "7b52f966-f86a-4396-9167-8b5d7ff9de20"
      },
      "source": [
        "score = model.evaluate(x_testcnn_encoded, y_test_encoded, verbose=0)\n",
        "print('Test loss:', score[0])\n",
        "print('Test accuracy:', score[1])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test loss: 0.11189327717988522\n",
            "Test accuracy: 0.9610389602648748\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RZIEcUhF2KJB",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 722
        },
        "outputId": "b0cad819-2863-4237-f90b-c91f2115a55f"
      },
      "source": [
        "predictions = model.predict_classes(x_testcnn_encoded)\n",
        "predictions\n",
        "y_test_encoded\n",
        "new_Ytest_encoded = y_test_encoded.astype(int)\n",
        "new_Ytest_encoded"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([3, 0, 3, 3, 2, 1, 4, 0, 4, 3, 5, 5, 3, 3, 3, 6, 0, 6, 5, 3, 4, 0,\n",
              "       6, 4, 0, 6, 5, 1, 0, 4, 1, 6, 0, 0, 3, 6, 5, 1, 0, 6, 2, 0, 3, 5,\n",
              "       6, 6, 0, 4, 4, 0, 1, 1, 1, 6, 3, 3, 2, 4, 2, 1, 3, 4, 6, 4, 5, 4,\n",
              "       0, 1, 4, 3, 3, 6, 1, 5, 0, 6, 4, 6, 0, 2, 0, 2, 0, 4, 6, 1, 1, 5,\n",
              "       6, 6, 5, 0, 3, 4, 0, 6, 1, 3, 3, 3, 3, 4, 5, 5, 4, 0, 4, 2, 0, 6,\n",
              "       1, 1, 4, 2, 0, 6, 0, 2, 3, 6, 3, 6, 4, 0, 3, 4, 2, 4, 0, 5, 3, 6,\n",
              "       5, 5, 5, 2, 3, 2, 3, 1, 2, 3, 4, 4, 2, 4, 4, 2, 1, 3, 6, 3, 4, 4,\n",
              "       5, 5, 0, 0, 3, 1, 1, 6, 4, 5, 1, 1, 1, 0, 2, 5, 0, 1, 1, 6, 1, 3,\n",
              "       4, 2, 5, 2, 1, 3, 2, 3, 4, 6, 6, 1, 4, 5, 6, 0, 5, 1, 6, 5, 4, 4,\n",
              "       6, 5, 3, 3, 6, 6, 0, 6, 2, 6, 3, 4, 2, 0, 3, 4, 6, 2, 2, 0, 4, 4,\n",
              "       6, 5, 2, 1, 3, 4, 0, 0, 3, 6, 4, 4, 2, 3, 2, 5, 6, 2, 3, 5, 4, 1,\n",
              "       0, 2, 1, 0, 1, 1, 4, 4, 3, 3, 2, 2, 3, 5, 3, 0, 6, 1, 3, 4, 5, 0,\n",
              "       0, 6, 5, 4, 3, 0, 3, 1, 4, 6, 0, 6, 2, 6, 1, 4, 4, 0, 4, 2, 2, 5,\n",
              "       2, 2, 6, 4, 5, 5, 3, 4, 4, 1, 4, 4, 2, 1, 4, 0, 3, 2, 4, 2, 3, 0,\n",
              "       1, 4, 5, 4, 3, 6, 3, 1, 5, 1, 4, 3, 2, 3, 2, 6, 4, 5, 6, 3, 2, 4,\n",
              "       3, 5, 6, 3, 5, 4, 2, 6, 0, 5, 5, 2, 0, 6, 5, 5, 6, 2, 1, 2, 4, 4,\n",
              "       4, 5, 6, 0, 6, 2, 1, 3, 6, 5, 4, 1, 0, 4, 3, 3, 5, 1, 0, 0, 5, 2,\n",
              "       2, 5, 1, 1, 4, 2, 1, 3, 4, 2, 3, 0, 6, 5, 4, 6, 2, 1, 1, 3, 1, 6,\n",
              "       2, 2, 0, 5, 5, 4, 3, 1, 1, 0, 2, 5, 0, 0, 6, 4, 5, 3, 6, 5, 4, 6,\n",
              "       2, 2, 4, 3, 6, 6, 6, 4, 6, 5, 5, 5, 4, 3, 2, 2, 0, 2, 0, 2, 1, 6,\n",
              "       0, 6, 6, 1, 0, 0, 5, 6, 5, 5, 0, 4, 5, 2, 2, 4, 6, 0, 5, 3, 6, 6,\n",
              "       3, 3, 0, 5, 6, 3, 3, 1, 1, 3, 4, 5, 6, 3, 5, 6, 5, 4, 4, 6, 2, 1,\n",
              "       0, 4, 0, 2, 1, 5, 1, 6, 5, 4, 5, 4, 2, 6, 5, 2, 5, 5, 4, 0, 1, 3,\n",
              "       4, 1, 4, 4, 5, 6, 3, 6, 6, 3, 0, 5, 4, 0, 1, 2, 4, 6, 0, 3, 1, 5,\n",
              "       3, 3, 0, 6, 2, 3, 6, 1, 0, 2, 1, 6, 1, 5, 5, 2, 2, 2, 2, 5, 1, 2,\n",
              "       1, 3, 5, 3, 3, 0, 3, 3, 0, 4, 6, 1, 2, 1, 0, 1, 5, 6, 1, 3, 2, 5,\n",
              "       0, 1, 3, 3, 5, 5, 4, 2, 1, 2, 0, 5, 3, 3, 6, 6, 6, 4, 6, 1, 1, 4,\n",
              "       1, 2, 5, 1, 2, 4, 5, 1, 2, 5, 0, 6, 3, 3, 5, 0, 5, 6, 3, 3, 4, 1,\n",
              "       4, 4, 2, 1, 5, 1, 0, 5, 5, 0, 1, 1, 0, 4, 2, 3, 3, 6, 5, 0, 3, 4,\n",
              "       1, 2, 0, 3, 2, 0, 6, 2, 1, 0, 5, 3, 0, 0, 6, 4, 3, 1, 5, 3, 0, 1,\n",
              "       0, 4, 0, 6, 6, 2, 5, 6, 6, 2, 2, 0, 6, 3, 1, 0, 2, 4, 3, 0, 0, 3,\n",
              "       1, 4, 6, 1, 1, 5, 0, 6, 5, 4, 1, 2, 2, 0, 6, 0, 2, 2, 5, 0, 2, 3,\n",
              "       0, 3, 0, 0, 5, 1, 4, 2, 1, 3, 5, 0, 2, 5, 3, 4, 3, 3, 1, 4, 5, 3,\n",
              "       3, 5, 3, 6, 6, 6, 4, 0, 5, 5, 3, 3, 3, 3, 1, 1, 6, 0, 4, 6, 3, 4,\n",
              "       1, 4, 0, 0, 1, 6, 1, 4, 4, 1, 5, 1, 0, 1, 2, 3, 2, 3, 1, 0, 6, 5,\n",
              "       1, 3, 6, 2, 5, 3, 4, 5, 6, 2, 6, 3, 2, 1, 5, 6, 0, 6, 4, 2, 4, 2,\n",
              "       0, 5, 3, 5, 1, 4, 3, 6, 6, 6, 6, 0, 4, 6, 0, 2, 2, 2, 5, 5, 6, 4,\n",
              "       5, 5, 6, 1, 4, 2, 0, 0, 0, 2, 3, 5, 4, 2, 1, 3, 2, 4, 4, 0, 2, 0,\n",
              "       1, 6, 5, 2, 5, 0, 1, 0, 5, 2, 2, 4, 1, 4, 3, 4, 5, 1, 5, 4, 1, 0,\n",
              "       5, 4, 6, 4, 1, 4, 1, 4, 2, 1, 2, 2, 4, 2, 4, 5, 0, 1, 2, 5, 2, 3,\n",
              "       2, 0, 5, 1, 0, 5, 1, 5, 5, 3, 2, 6, 6, 2, 2, 5, 2, 5, 4, 4, 3, 1,\n",
              "       1, 4, 3, 5, 4, 6, 0, 0, 0, 5, 1, 2, 0, 3, 2, 6, 2, 3, 2, 5, 6, 3])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 124
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vq97wtU22YqI",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 252
        },
        "outputId": "d37002a4-f9cd-4d33-b86e-af7351b81412"
      },
      "source": [
        "from sklearn.metrics import classification_report\n",
        "report = classification_report(new_Ytest_encoded, predictions)\n",
        "print(report)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.98      0.99      0.99       128\n",
            "           1       0.98      0.98      0.98       125\n",
            "           2       0.95      0.96      0.96       130\n",
            "           3       0.99      0.96      0.97       137\n",
            "           4       0.95      0.97      0.96       139\n",
            "           5       0.97      0.91      0.94       135\n",
            "           6       0.91      0.95      0.93       130\n",
            "\n",
            "    accuracy                           0.96       924\n",
            "   macro avg       0.96      0.96      0.96       924\n",
            "weighted avg       0.96      0.96      0.96       924\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mPEAY888-ieY",
        "colab_type": "text"
      },
      "source": [
        "PLOTTING THE CONFUSION MATRIX"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CwWE3ilP9QbI",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 134
        },
        "outputId": "da0c17cb-dd9f-421c-a825-a8146eee2a31"
      },
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "matrix = confusion_matrix(new_Ytest_encoded, predictions)\n",
        "print (matrix)\n",
        "\n",
        "# 0 = angry, 1 = disgust, 2 = fear, 3 = happy, 4 = neutral, 5 = surprise, 6 = sad"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[127   0   0   0   1   0   0]\n",
            " [  0 123   0   0   0   2   0]\n",
            " [  0   0 125   0   1   0   4]\n",
            " [  0   0   0 132   5   0   0]\n",
            " [  2   0   0   2 135   0   0]\n",
            " [  0   3   1   0   0 123   8]\n",
            " [  0   0   5   0   0   2 123]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cYq4-jeg-fur",
        "colab_type": "text"
      },
      "source": [
        "SAVE THE MODEL"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dYxQqcNq-Bjh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "0e3c644c-8658-462a-fc56-f940e909b112"
      },
      "source": [
        "model_name = 'Emotion_Voice_Detection_Model_TORONTO.h5'\n",
        "save_dir = '/content/drive/My Drive/Toronto_model'\n",
        "# Save model and weights\n",
        "if not os.path.isdir(save_dir):\n",
        "    os.makedirs(save_dir)\n",
        "model_path = os.path.join(save_dir, model_name)\n",
        "model.save(model_path)\n",
        "print('Saved trained model at %s ' % model_path)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Saved trained model at /content/drive/My Drive/Toronto_model/Emotion_Voice_Detection_Model_TORONTO.h5 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VTlyAvF6-MaI",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 403
        },
        "outputId": "4931e5ef-f60f-4557-e103-8e50d11a3029"
      },
      "source": [
        "loaded_model = keras.models.load_model('/content/drive/My Drive/Toronto_model/Emotion_Voice_Detection_Model_TORONTO.h5')\n",
        "loaded_model.summary()\n",
        "loss, acc = loaded_model.evaluate(x_testcnn_encoded, y_test_encoded)\n",
        "print(\"Restored model, accuracy: {:5.2f}%\".format(100*acc))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv1d_24 (Conv1D)           (None, 128, 32)           192       \n",
            "_________________________________________________________________\n",
            "max_pooling1d_14 (MaxPooling (None, 16, 32)            0         \n",
            "_________________________________________________________________\n",
            "dropout_25 (Dropout)         (None, 16, 32)            0         \n",
            "_________________________________________________________________\n",
            "flatten_14 (Flatten)         (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_76 (Dense)             (None, 128)               65664     \n",
            "_________________________________________________________________\n",
            "dropout_26 (Dropout)         (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_77 (Dense)             (None, 7)                 903       \n",
            "=================================================================\n",
            "Total params: 66,759\n",
            "Trainable params: 66,759\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "924/924 [==============================] - 1s 1ms/step\n",
            "Restored model, accuracy: 96.10%\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}