nikhilnanda21/medium_oversample.ipynb Secret

## medium_oversample.ipynb
{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "medium_oversample.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPbWh0UbmjMwjPeT5qqzfVd",
      "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/gist/nikhilnanda21/9ef9e78541a38962e73f0f9e1472397c/medium_oversample.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7jT32LHRgn01",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.model_selection import train_test_split\n",
        "import json"
      ],
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rUQusXtRgvRn",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 187
        },
        "outputId": "c2c164fc-a83b-49f5-ffdf-bc0be639af9e"
      },
      "source": [
        "df = pd.read_csv('/content/fer2013/fer2013/fer2013.csv')\n",
        "df.info()"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 35887 entries, 0 to 35886\n",
            "Data columns (total 3 columns):\n",
            " #   Column   Non-Null Count  Dtype \n",
            "---  ------   --------------  ----- \n",
            " 0   emotion  35887 non-null  int64 \n",
            " 1   pixels   35887 non-null  object\n",
            " 2   Usage    35887 non-null  object\n",
            "dtypes: int64(1), object(2)\n",
            "memory usage: 841.2+ KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SLgBu91Lg3iK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "1fa91fff-1aab-44aa-bf50-b9295935a2ad"
      },
      "source": [
        "sns.countplot(df.emotion)\n",
        "plt.xlabel('class')\n",
        "plt.title('Imbalance in dataset')\n",
        "plt.show()"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mGXNaXS3g7Hp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df_train, df_test = train_test_split(df, test_size=0.3, shuffle=True, stratify=df['emotion'])"
      ],
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cYu3xdxDiIDF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "46fa6cbf-15f7-4d4e-a90f-9297c5d67ca6"
      },
      "source": [
        "sns.countplot(df_train.emotion)\n",
        "plt.xlabel('class')\n",
        "plt.title('Imbalance in Training set')\n",
        "plt.show()"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8o-KDqWDiNkR",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "cellView": "code",
        "outputId": "84fb13ed-4a5f-4857-e3ab-10aa1cffe6ac"
      },
      "source": [
        "#@title Default title text\n",
        "sns.countplot(df.emotion)\n",
        "plt.xlabel('class')\n",
        "plt.title('Imbalance in Test set')\n",
        "plt.show()"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAWdklEQVR4nO3de7RdZX3u8e8jAS2IXFMKIRKq1BasFZsCSqseUERF4/Cogx5EtBzRMdDqOXq89oi32NpqLV5PqaCAF0TQI16qckDRWkESoCpEaopoErlEAgIqaPB3/lhv7DLs8K6NWXvtnXw/Y+yx53zfOd/5WzsZ+9nznXPNlapCkqR7cp9JFyBJmv0MC0lSl2EhSeoyLCRJXYaFJKnLsJAkdRkWmnOSfDDJm+/lvq9P8qHNXdO9leTPklw96TqkHsNCE5Hk2iSPm3Qdk1ZVX62qh0x3vySvSXJ7+7ojyV1D61fei/Eem2T1dPcbcWz/rbcAhoU0B1XVW6rq/lV1f+CFwNc3rFfVAZOuT1sew0ITl+S5Sb6W5B1JbklyTZJHtfZVSW5MctxGu+2e5PwktyW5KMk+Q+Od3Pa7NcnyJH92D8f+eJLrk/w4yVeSHDDU98Ek70ny2XacS5I8aKj/gFbDuiQ3JHlNa79Pklcl+Y8kNyU5O8mumzj+r/1F3/4Kf3mSb7aaPpbkftP8ef7+UF1XJ3nWUN+TklzVXs+adqwdgH8G9ho6O9lrinHvtu9Q31FJrmj/fv+a5GGt/UzggcCn27ivmM5r0exhWGi2OBj4JrAb8BHgLOBPgAcDzwbeneT+Q9sfA7wJ2B24AvjwUN+lwMOBXdtYH7+HX7j/DOwH/DZw2UbjABwNvAHYBVgJLAVIsiPw/4DPA3u1Oi9o+7wYeBrwmNZ3M/CekX4KA88CjgT2BR4GPHfUHdsv/vMZvO7fbvW/N8n+bZNTgRdU1Y7AQ4ELq+onwBOBHw6dnfxwiuHvtm875oHAacALGPz7/SNwXpL7VtWxwA+Ap7Rx/3YaPwfNIoaFZovvVdUHquou4GPAQuCNVXVnVX0R+DmDX8gbfLaqvlJVdwKvBR6ZZCFAVX2oqm6qqvVV9XbgvsCU1wWq6rSquq2N83rgj5LsNLTJJ6vqG1W1nkGQPLy1HwVcX1Vvr6o72hiXtL4XAq+tqtVD4z4jybwRfxbvrKofVtU64NNDxxzFUcC17We5vqouB84Fntn6fwHsn+QBVXVzVV02jbE3te8JwD9W1SVVdVdVnQ7cCRwyjbE1yxkWmi1uGFr+GUBVbdw2fGaxasNCVd0OrGPwVzxtamVFm8a5BdiJwRnIr0myTZK/adNFtwLXtq7hba8fWv7pUA0Lgf/YxGvZB/hkm5K5BVgB3AXssYntN7apY45iH+DgDcduxz8G+J3W/1+BJwHfb9N3j5zG2Jvadx/gZRsdcyHt30NbhlH/0pFmm4UbFtr01K7AD9v1iVcAhwNXVtUvk9wMZIox/huwBHgcg6DYicGU0VTbbmwVgymeTfX9RVV9bbSXslmtAi6qqsdP1VlVlwJLkmwLvAg4m8HPsvv46XvYdxWwtKqWbmrXab8KzTqeWWiuelKSP02yHYNrFxdX1SpgR2A9sBaYl+R1wAM2McaODKZLbgK2B94yjeN/BtgzyUuT3DfJjkkObn3/B1i64aJ7kvlJlkz3Bd5LnwF+L8mxSbZtX3+S5A+SbJfkmCQ7VdUvgFuBX7b9bgB222gK7lc6+/4T8MIkB2dghyRPbtd1Noz9u2N7xZoRhoXmqo8AJzGYfvpjBhfBAb7A4KLzvwPfB+5gaMpqI2e0bdYAVwEXj3rwqroNeDzwFAbTRt8F/kvrPhk4D/hiktvauAdPNc7m1uo6gsFZzw9bbW9lcN0G4Fjg2jbt9kIGU1RU1XeAjwLXtKmkqaaQNrXvMuD5wLsZnJmt5Ncvyv818Fdt3JejOSl++JEkqcczC0lSl2EhSeoyLCRJXYaFJKlri3yfxe67716LFi2adBmSNKcsX778R1U1f6q+LTIsFi1axLJlyyZdhiTNKUm+v6k+p6EkSV2GhSSpy7CQJHUZFpKkLsNCktRlWEiSugwLSVKXYSFJ6jIsJEldW+Q7uKXN7aJHP2bSJUzpMV+5aNIlaCvhmYUkqcuwkCR1GRaSpC7DQpLUZVhIkroMC0lSl2EhSeoyLCRJXYaFJKnLsJAkdRkWkqQuw0KS1GVYSJK6DAtJUpdhIUnqMiwkSV2GhSSpy7CQJHWNNSyS/I8kVyb5dpKPJrlfkn2TXJJkZZKPJdmubXvftr6y9S8aGufVrf3qJE8YZ82SpLsbW1gkWQD8JbC4qh4KbAMcDbwVeEdVPRi4GTi+7XI8cHNrf0fbjiT7t/0OAI4E3ptkm3HVLUm6u3FPQ80DfivJPGB74DrgMOCc1n868LS2vKSt0/oPT5LWflZV3VlV3wNWAgeNuW5J0pCxhUVVrQHeBvyAQUj8GFgO3FJV69tmq4EFbXkBsKrtu75tv9tw+xT7/EqSE5IsS7Js7dq1m/8FSdJWbJzTULswOCvYF9gL2IHBNNJYVNUpVbW4qhbPnz9/XIeRpK3SOKehHgd8r6rWVtUvgE8AhwI7t2kpgL2BNW15DbAQoPXvBNw03D7FPpKkGTDOsPgBcEiS7du1h8OBq4AvAc9o2xwHfKotn9fWaf0XVlW19qPb3VL7AvsB3xhj3ZKkjczrb3LvVNUlSc4BLgPWA5cDpwCfBc5K8ubWdmrb5VTgzCQrgXUM7oCiqq5McjaDoFkPnFhVd42rbknS3Y0tLACq6iTgpI2ar2GKu5mq6g7gmZsYZymwdLMXKEkaie/gliR1GRaSpC7DQpLUZVhIkroMC0lSl2EhSeoyLCRJXYaFJKnLsJAkdRkWkqQuw0KS1GVYSJK6DAtJUpdhIUnqMiwkSV2GhSSpy7CQJHUZFpKkLsNCktRlWEiSugwLSVKXYSFJ6jIsJEldhoUkqcuwkCR1GRaSpC7DQpLUZVhIkroMC0lSl2EhSeoyLCRJXYaFJKnLsJAkdRkWkqQuw0KS1GVYSJK65k26AEnakq1YeuGkS5jSH7z2sGltP9YziyQ7JzknyXeSrEjyyCS7Jjk/yXfb913atknyziQrk3wzySOGxjmubf/dJMeNs2ZJ0t2NexrqZODzVfX7wB8BK4BXARdU1X7ABW0d4InAfu3rBOB9AEl2BU4CDgYOAk7aEDCSpJkxtmmoJDsBjwaeC1BVPwd+nmQJ8Ni22enAl4FXAkuAM6qqgIvbWcmebdvzq2pdG/d84Ejgo+OqXdqSvPtln550CZv0orc/ZdIlaETjPLPYF1gLfCDJ5Unen2QHYI+quq5tcz2wR1teAKwa2n91a9tUuyRphowzLOYBjwDeV1UHAj/hP6ecAGhnEbU5DpbkhCTLkixbu3bt5hhSktSMMyxWA6ur6pK2fg6D8LihTS/Rvt/Y+tcAC4f237u1bar911TVKVW1uKoWz58/f7O+EEna2o0tLKrqemBVkoe0psOBq4DzgA13NB0HfKotnwc8p90VdQjw4zZd9QXgiCS7tAvbR7Q2SdIMGff7LF4MfDjJdsA1wPMYBNTZSY4Hvg88q237OeBJwErgp21bqmpdkjcBl7bt3rjhYrckaWaMNSyq6gpg8RRdh0+xbQEnbmKc04DTNm91kqRR+bgPSVKXYSFJ6jIsJEldhoUkqcuwkCR1GRaSpC7DQpLUZVhIkroMC0lSl2EhSeoyLCRJXYaFJKnLsJAkdY0UFkkuGKVNkrRlusdHlCe5H7A9sHv74KG0rgfg52BL0laj93kWLwBeCuwFLOc/w+JW4N1jrEtbmEPfdeikS5jS1178tUmXIM0J9xgWVXUycHKSF1fVu2aoJknSLDPSJ+VV1buSPApYNLxPVZ0xprokSbPISGGR5EzgQcAVwF2tuQDDQpK2AqN+BvdiYP/2Odlz0h//r9mZa8v/7jmTLkGSukZ9n8W3gd8ZZyGSpNlr1DOL3YGrknwDuHNDY1U9dSxVSZJmlVHD4vXjLEKSNLuNejfUReMuRJI0e416N9RtDO5+AtgO2Bb4SVU9YFyFSZJmj1HPLHbcsJwkwBLgkHEVJUmaXab91Nka+L/AE8ZQjyRpFhp1GurpQ6v3YfC+izvGUpEkadYZ9W6opwwtrweuZTAVJUnaCox6zeJ54y5EkjR7jfrhR3sn+WSSG9vXuUn2HndxkqTZYdQL3B8AzmPwuRZ7AZ9ubZKkrcCoYTG/qj5QVevb1weB+WOsS5I0i4waFjcleXaSbdrXs4GbxlmYJGn2GDUs/gJ4FnA9cB3wDOC5Y6pJkjTLjHrr7BuB46rqZoAkuwJvYxAikqQt3KhnFg/bEBQAVbUOOHA8JUmSZptRw+I+SXbZsNLOLEZ99/c2SS5P8pm2vm+SS5KsTPKxJNu19vu29ZWtf9HQGK9u7Vcn8TEjkjTDRg2LtwNfT/KmJG8C/hX42xH3fQmwYmj9rcA7qurBwM3A8a39eODm1v6Oth1J9geOBg4AjgTem2SbEY8tSdoMRgqLqjoDeDpwQ/t6elWd2duvvXHvycD723qAw4Bz2ianA09ry0vaOq3/8KEn3J5VVXdW1feAlcBBo9QtSdo8Rr3ATVVdBVw1zfH/AXgFsOER57sBt1TV+ra+GljQlhcAq9qx1if5cdt+AXDx0JjD+0iSZsDIYTFdSY4Cbqyq5UkeO67jDB3vBOAEgAc+8IHjPpykGbL02c+YdAmb9NoPndPfaAsx7c+zmIZDgacmuRY4i8H008nAzkk2hNTewJq2vAZYCND6d2Lwxr9ftU+xz69U1SlVtbiqFs+f75vLJWlzGltYVNWrq2rvqlrE4AL1hVV1DPAlBm/qAzgO+FRbPq+t0/ovrKpq7Ue3u6X2BfYDvjGuuiVJdze2aah78ErgrCRvBi4HTm3tpwJnJlkJrGMQMFTVlUnOZnC9ZD1wYlXdNfNlS9LWa0bCoqq+DHy5LV/DFHczVdUdwDM3sf9SYOn4KpQk3ZNxXrOQJG0hDAtJUpdhIUnqMiwkSV2GhSSpy7CQJHUZFpKkLsNCktRlWEiSugwLSVKXYSFJ6jIsJEldhoUkqcuwkCR1GRaSpC7DQpLUZVhIkroMC0lSl2EhSeoyLCRJXYaFJKnLsJAkdRkWkqQuw0KS1GVYSJK6DAtJUpdhIUnqMiwkSV2GhSSpy7CQJHUZFpKkLsNCktRlWEiSugwLSVKXYSFJ6jIsJEldhoUkqcuwkCR1jS0skixM8qUkVyW5MslLWvuuSc5P8t32fZfWniTvTLIyyTeTPGJorOPa9t9Ncty4apYkTW2cZxbrgZdV1f7AIcCJSfYHXgVcUFX7ARe0dYAnAvu1rxOA98EgXICTgIOBg4CTNgSMJGlmjC0squq6qrqsLd8GrAAWAEuA09tmpwNPa8tLgDNq4GJg5yR7Ak8Azq+qdVV1M3A+cOS46pYk3d2MXLNIsgg4ELgE2KOqrmtd1wN7tOUFwKqh3Va3tk21b3yME5IsS7Js7dq1m7V+SdrajT0sktwfOBd4aVXdOtxXVQXU5jhOVZ1SVYuravH8+fM3x5CSpGasYZFkWwZB8eGq+kRrvqFNL9G+39ja1wALh3bfu7Vtql2SNEPGeTdUgFOBFVX190Nd5wEb7mg6DvjUUPtz2l1RhwA/btNVXwCOSLJLu7B9RGuTJM2QeWMc+1DgWOBbSa5oba8B/gY4O8nxwPeBZ7W+zwFPAlYCPwWeB1BV65K8Cbi0bffGqlo3xrolSRsZW1hU1b8A2UT34VNsX8CJmxjrNOC0zVedJGk6fAe3JKnLsJAkdRkWkqQuw0KS1GVYSJK6DAtJUpdhIUnqMiwkSV2GhSSpy7CQJHUZFpKkLsNCktRlWEiSugwLSVKXYSFJ6jIsJEldhoUkqcuwkCR1GRaSpC7DQpLUZVhIkroMC0lSl2EhSeoyLCRJXYaFJKnLsJAkdRkWkqSueZMuQKP5wRv/cNIlTOmBr/vWpEuQNAM8s5AkdRkWkqQuw0KS1GVYSJK6DAtJUpdhIUnqMiwkSV2GhSSpy7CQJHXNmbBIcmSSq5OsTPKqSdcjSVuTOREWSbYB3gM8Edgf+PMk+0+2KknaesyJsAAOAlZW1TVV9XPgLGDJhGuSpK1GqmrSNXQleQZwZFX997Z+LHBwVb1oaJsTgBPa6kOAq8dY0u7Aj8Y4/rhZ/2RZ/+TM5dph/PXvU1Xzp+rYYp46W1WnAKfMxLGSLKuqxTNxrHGw/smy/smZy7XDZOufK9NQa4CFQ+t7tzZJ0gyYK2FxKbBfkn2TbAccDZw34ZokaasxJ6ahqmp9khcBXwC2AU6rqisnWNKMTHeNkfVPlvVPzlyuHSZY/5y4wC1Jmqy5Mg0lSZogw0KS1GVYTNNcfuxIktOS3Jjk25OuZbqSLEzypSRXJbkyyUsmXdN0JLlfkm8k+bdW/xsmXdO9kWSbJJcn+cyka5muJNcm+VaSK5Ism3Q905Vk5yTnJPlOkhVJHjmjx/eaxejaY0f+HXg8sJrBXVp/XlVXTbSwESV5NHA7cEZVPXTS9UxHkj2BPavqsiQ7AsuBp82hn32AHarq9iTbAv8CvKSqLp5wadOS5H8Ci4EHVNVRk65nOpJcCyyuqjn5prwkpwNfrar3t7tCt6+qW2bq+J5ZTM+cfuxIVX0FWDfpOu6Nqrquqi5ry7cBK4AFk61qdDVwe1vdtn3Nqb/UkuwNPBl4/6Rr2dok2Ql4NHAqQFX9fCaDAgyL6VoArBpaX80c+oW1pUiyCDgQuGSylUxPm8K5ArgROL+q5lT9wD8ArwB+OelC7qUCvphkeXs80FyyL7AW+ECbBnx/kh1msgDDQnNKkvsD5wIvrapbJ13PdFTVXVX1cAZPIDgoyZyZCkxyFHBjVS2fdC2/gT+tqkcweHr1iW1adq6YBzwCeF9VHQj8BJjRa6aGxfT42JEJanP95wIfrqpPTLqee6tNH3wJOHLStUzDocBT27z/WcBhST402ZKmp6rWtO83Ap9kMK08V6wGVg+djZ7DIDxmjGExPT52ZELaBeJTgRVV9feTrme6ksxPsnNb/i0GN0l8Z7JVja6qXl1Ve1fVIgb/7y+sqmdPuKyRJdmh3RhBm745ApgzdwVW1fXAqiQPaU2HAzN6c8eceNzHbDELHzsyLUk+CjwW2D3JauCkqjp1slWN7FDgWOBbbd4f4DVV9bkJ1jQdewKntzvq7gOcXVVz7vbTOWwP4JODvzmYB3ykqj4/2ZKm7cXAh9sfqtcAz5vJg3vrrCSpy2koSVKXYSFJ6jIsJEldhoUkqcuwkCR1GRbSGCR5fZKXT7oOaXMxLCRJXYaFtBkkeU6Sb7bPqzhzo77nJ7m09Z2bZPvW/swk327tX2ltB7TPvbiijbffJF6PtDHflCf9hpIcwOBZQ4+qqh8l2RX4S+D2qnpbkt2q6qa27ZuBG6rqXUm+BRxZVWuS7FxVtyR5F3BxVW14p+42VfWzSb02aQPPLKTf3GHAxzd8qE5VbfyZIQ9N8tUWDscAB7T2rwEfTPJ8Bo+PAfg68JokrwT2MSg0WxgW0vh9EHhRVf0h8AbgfgBV9ULgrxg8yXh5OwP5CPBU4GfA55IcNpmSpV9nWEi/uQuBZybZDaBNQw3bEbiuPWL9mA2NSR5UVZdU1esYfLDNwiS/C1xTVe8EPgU8bEZegdThU2el31BVXZlkKXBRkruAy4Frhzb53ww+1W9t+75ja/+7dgE7wAXAvwGvBI5N8gvgeuAtM/IipA4vcEuSupyGkiR1GRaSpC7DQpLUZVhIkroMC0lSl2EhSeoyLCRJXf8fBIyHAAnIhaYAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gqHo0NFbiQMt",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "4e65831c-7e6e-4b56-e791-338fd02b220d"
      },
      "source": [
        "class_count = json.loads(df_train['emotion'].value_counts().to_json())\n",
        "class_count"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'0': 3467, '1': 383, '2': 3585, '3': 6292, '4': 4254, '5': 2801, '6': 4338}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bcEYaoNnibe1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "f89c36c1-cd8e-4e7e-ea4a-95b7591dd700"
      },
      "source": [
        "max_class_value = max(class_count.values()) \n",
        "max_classes = [key for key in class_count if class_count[key] == max_class_value]\n",
        "max_class = max_classes[0]\n",
        "max_class_images = class_count[max_class]\n",
        "max_class, max_class_images"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('3', 6292)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qcQ4r7_Lks7b",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49
        },
        "outputId": "bd24fa40-0a55-4d1f-8b95-3db17bb3a677"
      },
      "source": [
        "df_train_over = pd.DataFrame(columns = df.columns)\n",
        "df_train_over"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>emotion</th>\n",
              "      <th>pixels</th>\n",
              "      <th>Usage</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "Empty DataFrame\n",
              "Columns: [emotion, pixels, Usage]\n",
              "Index: []"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qWgvHJm_iwJj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for i in range(len(class_count)):\n",
        "  df_one_class = df_train[df_train['emotion']==i]\n",
        "  df_one_class_over = df_one_class.sample(max_class_images, replace=True)\n",
        "  df_train_over = pd.concat([df_train_over, df_one_class_over], axis=0)"
      ],
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7emo8Z9LrmGA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 153
        },
        "outputId": "328c931e-019a-4ea3-bb33-05fd043385d7"
      },
      "source": [
        "df_train_over.emotion.value_counts()"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6    6292\n",
              "5    6292\n",
              "4    6292\n",
              "3    6292\n",
              "2    6292\n",
              "1    6292\n",
              "0    6292\n",
              "Name: emotion, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bs6v2pS8lElU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "dd0c2f61-c747-493e-e10c-06ddd665694a"
      },
      "source": [
        "sns.countplot(df_train_over.emotion)\n",
        "plt.xlabel('class')\n",
        "plt.title('Oversampled Train set')\n",
        "plt.show()"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FwMWw3zulQik",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 46,
      "outputs": []
    }
  ]
}