Session_7/(2) Linear, Lasso,  Ridge and Elastic Regressions Notebook.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Linear, Lasso, Ridge and Elastic Regressions.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import math\n",
        "import matplotlib.pyplot as plt\n",
        "from scipy.interpolate import make_interp_spline\n",
        "from sklearn.datasets import load_boston, fetch_california_housing\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.linear_model import LinearRegression, Ridge, RidgeCV, Lasso, LassoCV, ElasticNet, ElasticNetCV\n",
        "from sklearn import metrics\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")"
      ],
      "metadata": {
        "id": "xLQU8vHvFPA-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# url = 'http://archive.ics.uci.edu/ml/machine-learning-databases/communities/communities.data'\n",
        "\n",
        "# df = pd.read_csv(url, header = None, na_values = ['?'])\n",
        "\n",
        "# df.drop([0, 1, 2, 3, 4], axis = 1, inplace = True)\n",
        "# df.dropna(inplace = True)\n",
        "\n",
        "# X = df.drop(127, axis = 1, inplace = False)\n",
        "# y = df[127]\n",
        "\n",
        "# X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3, random_state = 1)"
      ],
      "metadata": {
        "id": "FRjVmxULyPl4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# df = load_boston()\n",
        "\n",
        "# housing = pd.DataFrame(df.data)\n",
        "\n",
        "# housing.columns = df.feature_names\n",
        "# housing['Price'] = df.target\n",
        "\n",
        "# X = housing.drop('Price', axis = 1, inplace = False)\n",
        "# y = housing['Price']\n",
        "\n",
        "# print(housing)\n",
        "\n",
        "# X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.4, random_state = 1, shuffle = True)"
      ],
      "metadata": {
        "id": "EuAdOFAzFgsk"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df = fetch_california_housing()\n",
        "\n",
        "housing = pd.DataFrame(df.data)\n",
        "\n",
        "housing.columns = df.feature_names\n",
        "housing['Price'] = df.target\n",
        "\n",
        "X = housing.drop('Price', axis = 1, inplace = False)\n",
        "y = housing['Price']\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.8, random_state = 1, shuffle = True)"
      ],
      "metadata": {
        "id": "dQwBSrpTJS9Y"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Linear Regression\n",
        "\n"
      ],
      "metadata": {
        "id": "jkWOorpqFDAU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "linear_regression = LinearRegression()\n",
        "\n",
        "linear_regression.fit(X_train, y_train)\n",
        "\n",
        "LinearRegression(copy_X = True, fit_intercept = True, n_jobs = None, normalize = False)\n",
        "\n",
        "print(\"Coefficients :\", linear_regression.coef_, \"\\n\")\n",
        "\n",
        "train_pred_linear = linear_regression.predict(X_train)\n",
        "test_pred_linear = linear_regression.predict(X_test)\n",
        "\n",
        "print(\"Training Error on Linear Regression Model :\", metrics.mean_squared_error(y_train, train_pred_linear), \"\\n\")\n",
        "print(\"Generalisation Error of Linear Regression Model :\", metrics.mean_squared_error(y_test, test_pred_linear))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "n3DcxJ3aFErW",
        "outputId": "f619f40f-a09d-43ea-92de-8c4a44ec5161"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Coefficients : [ 3.95756322e-01  1.09421458e-02 -1.43702376e-02  1.63555930e-01\n",
            "  1.36694668e-05 -2.51888795e-01 -4.33561886e-01 -4.16993317e-01] \n",
            "\n",
            "Training Error on Linear Regression Model : 0.4864539687571813 \n",
            "\n",
            "Generalisation Error of Linear Regression Model : 8.844120921880256\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Lasso Regression(L1 Regularisation)"
      ],
      "metadata": {
        "id": "TWZ_Oq0qEyDv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "lasso_regression = Lasso(alpha = 0.0005, normalize = True)\n",
        "\n",
        "lasso_regression.fit(X_train, y_train)\n",
        "\n",
        "print(\"Coefficients :\", lasso_regression.coef_, \"\\n\")\n",
        "\n",
        "train_pred_lasso = lasso_regression.predict(X_train)\n",
        "test_pred_lasso = lasso_regression.predict(X_test)\n",
        "\n",
        "print(\"Training Error on Lasso Regression Model :\", metrics.mean_squared_error(y_train, train_pred_lasso), \"\\n\")\n",
        "print(\"Generalisation Error of Lasso Regression Model :\", metrics.mean_squared_error(y_test, test_pred_lasso))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "577qyrMlEzzZ",
        "outputId": "4a9cc0de-35d0-4a0a-d7b2-d42290cbe6ae"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Coefficients : [ 0.39398091  0.01132389  0.          0.01214647  0.         -0.22097932\n",
            " -0.2217145  -0.19283625] \n",
            "\n",
            "Training Error on Lasso Regression Model : 0.518915834842168 \n",
            "\n",
            "Generalisation Error of Lasso Regression Model : 6.937543941227409\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "lasso_regression_cv = LassoCV(n_alphas = 1000, normalize = True, random_state = 1)\n",
        "\n",
        "lasso_regression_cv.fit(X_train, y_train)\n",
        "\n",
        "print(\"Coefficients :\", lasso_regression.coef_, \"\\n\")\n",
        "\n",
        "print(\"Chosen regularisation parameter :\", lasso_regression_cv.alpha_, \"\\n\")\n",
        "\n",
        "train_pred_lassocv = lasso_regression.predict(X_train)\n",
        "test_pred_lassocv = lasso_regression_cv.predict(X_test)\n",
        "\n",
        "print(\"Training Error of Lasso Regression Model :\", metrics.mean_squared_error(y_train, train_pred_lassocv), \"\\n\")\n",
        "print(\"Generalisation Error of Lasso Regression Model : \", metrics.mean_squared_error(y_test, test_pred_lassocv))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u6kjl3GpnTrA",
        "outputId": "be0ebb11-fcfd-4f83-a8d0-ddfcba9e8953"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Coefficients : [ 0.39398091  0.01132389  0.          0.01214647  0.         -0.22097932\n",
            " -0.2217145  -0.19283625] \n",
            "\n",
            "Chosen regularisation parameter : 0.00015920965170119252 \n",
            "\n",
            "Training Error of Lasso Regression Model : 0.518915834842168 \n",
            "\n",
            "Generalisation Error of Lasso Regression Model :  8.210627054413106\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def lasso(X_train, y_train, X_test, y_test, alpha):\n",
        "    lasso = Lasso(alpha = alpha, normalize = True)\n",
        "    lasso.fit(X_train, y_train)\n",
        "    return lasso.predict(X_test)\n",
        "\n",
        "alphas = np.array(lasso_regression_cv.alphas_)\n",
        "\n",
        "x = np.array(alphas)\n",
        "x_log = np.array([math.log(alpha, 10) for alpha in alphas])\n",
        "y = np.array([metrics.mean_squared_error(y_test, lasso(X_train, y_train, X_test, y_test, alpha)) for alpha in alphas])\n",
        "\n",
        "plt.plot(x, y)\n",
        "plt.title(\"Generalisation Error v/s Regularisation parameter\")\n",
        "plt.xlabel(\"Regularisation Parameter\")\n",
        "plt.ylabel(\"Generalisation Error\")\n",
        "plt.show()\n",
        "\n",
        "plt.plot(x_log, y)\n",
        "plt.title(\"Generalisation Error v/s Logarithm of Regularisation parameter\")\n",
        "plt.xlabel(\"log(Regularisation Parameter)\")\n",
        "plt.ylabel(\"Generalisation Error\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "iFQbGeL-Yw40",
        "outputId": "1da82d46-1780-4af3-d3b2-75697c0bb2f9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Ridge Regression(L2 Regularisation)"
      ],
      "metadata": {
        "id": "B5u9IPT4EzmJ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "s-lH3D79_fGC",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "91c0811c-6491-4a1b-bec0-7fb8574869c9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Coefficients : \n",
            " [ 3.76760328e-01  1.33376587e-02  3.60605818e-03  3.25448371e-02\n",
            "  2.19542323e-05 -2.38957591e-01 -1.96088622e-01 -1.65868238e-01] \n",
            "\n",
            "Training Error of Ridge Regression Model : 0.5263808713584214 \n",
            "\n",
            "Generalisation Error of Ridge Regression Model : 8.033483614037692\n"
          ]
        }
      ],
      "source": [
        "ridge_regression = Ridge(alpha = 0.1, normalize = True)\n",
        "\n",
        "ridge_regression.fit(X_train, y_train)\n",
        "\n",
        "print(\"Coefficients : \\n\", ridge_regression.coef_, \"\\n\")\n",
        "\n",
        "train_pred_ridge = ridge_regression.predict(X_train)\n",
        "test_pred_ridge = ridge_regression.predict(X_test)\n",
        "\n",
        "print(\"Training Error of Ridge Regression Model :\", metrics.mean_squared_error(y_train, train_pred_ridge), \"\\n\")\n",
        "print(\"Generalisation Error of Ridge Regression Model :\", metrics.mean_squared_error(y_test, test_pred_ridge))"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "alphas = [1e-5, 1e-4, 1e-3, 1e-2, 1e-1, 1, 2, 5, 10, 50, 100, 200, 500, 1000]\n",
        "\n",
        "ridge_regression_cv = RidgeCV(alphas = alphas, normalize = True, scoring = 'neg_mean_squared_error')\n",
        "\n",
        "ridge_regression_cv.fit(X_train, y_train)\n",
        "\n",
        "#RidgeCV automatically chooses the best regularisation parameter possible\n",
        "print(\"Chosen regularisation parameter :\", ridge_regression_cv.alpha_, \"\\n\")\n",
        "\n",
        "print(\"Coefficients : \\n\", ridge_regression_cv.coef_, \"\\n\")\n",
        "\n",
        "train_pred_ridgecv = ridge_regression_cv.predict(X_train)\n",
        "test_pred_ridgecv = ridge_regression_cv.predict(X_test)\n",
        "\n",
        "print(\"Training Error of Ridge Regression Model :\", metrics.mean_squared_error(y_train, train_pred_ridgecv))\n",
        "print(\"Generalisation Error of Ridge Regression Model :\", metrics.mean_squared_error(y_test, test_pred_ridgecv))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DwNs1pSdhmxR",
        "outputId": "55e42830-7c79-49b3-eb79-706340947429"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Chosen regularisation parameter : 0.01 \n",
            "\n",
            "Coefficients : \n",
            " [ 3.97481045e-01  1.17773842e-02 -1.28759883e-02  1.45496184e-01\n",
            "  1.70477924e-05 -2.51612770e-01 -3.80353957e-01 -3.60663140e-01] \n",
            "\n",
            "Training Error of Ridge Regression Model : 0.4881552574913633\n",
            "Generalisation Error of Ridge Regression Model : 8.822950246148862\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def ridge(X_train, y_train, X_test, y_test, alpha):\n",
        "    ridge = Ridge(alpha = alpha, normalize = True)\n",
        "    ridge.fit(X_train, y_train)\n",
        "    return ridge.predict(X_test)\n",
        "\n",
        "x_log = np.array([math.log(alpha, 10) for alpha in alphas])\n",
        "x = np.array(alphas)\n",
        "y = np.array([metrics.mean_squared_error(y_test, ridge(X_train, y_train, X_test, y_test, alpha)) for alpha in alphas])\n",
        "\n",
        "plt.plot(x, y)\n",
        "plt.title(\"Generalisation Error v/s Regularisation parameter\")\n",
        "plt.xlabel(\"Regularisation Parameter\")\n",
        "plt.ylabel(\"Generalisation Error\")\n",
        "plt.show()\n",
        "\n",
        "plt.plot(x_log, y)\n",
        "plt.title(\"Generalisation Error v/s Logarithm of Regularisation parameter\")\n",
        "plt.xlabel(\"log(Regularisation Parameter)\")\n",
        "plt.ylabel(\"Generalisation Error\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "rckNbUiqYg5Z",
        "outputId": "47d50e50-76dc-4c84-d400-117257a8a849"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "elastic_regression = ElasticNet(alpha = 0.001, l1_ratio = 0.3, normalize = True)\n",
        "\n",
        "elastic_regression.fit(X_train, y_train)\n",
        "\n",
        "print(\"Coefficients : \\n\", elastic_regression.coef_, \"\\n\")\n",
        "\n",
        "train_pred_elastic = elastic_regression.predict(X_train)\n",
        "test_pred_elastic = elastic_regression.predict(X_test)\n",
        "\n",
        "print(\"Training Error of Elastic Regression Model :\", metrics.mean_squared_error(y_train, train_pred_elastic), \"\\n\")\n",
        "print(\"Generalisation Error of Elastic Regression Model :\", metrics.mean_squared_error(y_test, test_pred_elastic))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gOPB5aue3WsY",
        "outputId": "8e66bfd3-48e3-4edd-f607-29763913f43d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Coefficients : \n",
            " [ 1.06566609e-01  2.38680373e-03  6.84439500e-03 -3.94822199e-03\n",
            " -2.51569274e-06 -6.49296803e-02 -2.04625089e-02 -4.28981486e-03] \n",
            "\n",
            "Training Error of Elastic Regression Model : 0.9948646996252336 \n",
            "\n",
            "Generalisation Error of Elastic Regression Model : 1.5502852179057978\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "elastic_regression_cv = ElasticNetCV(l1_ratio = 0.7, n_alphas = 100, normalize = True)\n",
        "\n",
        "elastic_regression_cv.fit(X_train, y_train)\n",
        "\n",
        "print(\"Chosen regularisation parameter :\", ridge_regression_cv.alpha_, \"\\n\")\n",
        "\n",
        "print(\"Coefficients : \\n\", elastic_regression.coef_, \"\\n\")\n",
        "\n",
        "train_pred_elasticcv = elastic_regression.predict(X_train)\n",
        "test_pred_elasticcv = elastic_regression.predict(X_test)\n",
        "\n",
        "print(\"Alphas :\\n\", elastic_regression_cv.alphas_, \"\\n\")\n",
        "\n",
        "print(\"Training Error of Elastic Regression Model :\", metrics.mean_squared_error(y_train, train_pred_elasticcv), \"\\n\")\n",
        "print(\"Generalisation Error of Elastic Regression Model :\", metrics.mean_squared_error(y_test, test_pred_elasticcv))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "st7ffJpK3XRH",
        "outputId": "93f3e6e7-7bd0-4c0a-9be2-1d785b290ff7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Chosen regularisation parameter : 0.01 \n",
            "\n",
            "Coefficients : \n",
            " [ 1.06566609e-01  2.38680373e-03  6.84439500e-03 -3.94822199e-03\n",
            " -2.51569274e-06 -6.49296803e-02 -2.04625089e-02 -4.28981486e-03] \n",
            "\n",
            "Alphas :\n",
            " [1.76100359e-02 1.64231784e-02 1.53163112e-02 1.42840430e-02\n",
            " 1.33213463e-02 1.24235322e-02 1.15862277e-02 1.08053547e-02\n",
            " 1.00771100e-02 9.39794650e-03 8.76455636e-03 8.17385460e-03\n",
            " 7.62296415e-03 7.10920188e-03 6.63006547e-03 6.18322125e-03\n",
            " 5.76649283e-03 5.37785051e-03 5.01540139e-03 4.67738012e-03\n",
            " 4.36214035e-03 4.06814669e-03 3.79396722e-03 3.53826653e-03\n",
            " 3.29979921e-03 3.07740378e-03 2.86999707e-03 2.67656887e-03\n",
            " 2.49617709e-03 2.32794311e-03 2.17104753e-03 2.02472620e-03\n",
            " 1.88826643e-03 1.76100359e-03 1.64231784e-03 1.53163112e-03\n",
            " 1.42840430e-03 1.33213463e-03 1.24235322e-03 1.15862277e-03\n",
            " 1.08053547e-03 1.00771100e-03 9.39794650e-04 8.76455636e-04\n",
            " 8.17385460e-04 7.62296415e-04 7.10920188e-04 6.63006547e-04\n",
            " 6.18322125e-04 5.76649283e-04 5.37785051e-04 5.01540139e-04\n",
            " 4.67738012e-04 4.36214035e-04 4.06814669e-04 3.79396722e-04\n",
            " 3.53826653e-04 3.29979921e-04 3.07740378e-04 2.86999707e-04\n",
            " 2.67656887e-04 2.49617709e-04 2.32794311e-04 2.17104753e-04\n",
            " 2.02472620e-04 1.88826643e-04 1.76100359e-04 1.64231784e-04\n",
            " 1.53163112e-04 1.42840430e-04 1.33213463e-04 1.24235322e-04\n",
            " 1.15862277e-04 1.08053547e-04 1.00771100e-04 9.39794650e-05\n",
            " 8.76455636e-05 8.17385460e-05 7.62296415e-05 7.10920188e-05\n",
            " 6.63006547e-05 6.18322125e-05 5.76649283e-05 5.37785051e-05\n",
            " 5.01540139e-05 4.67738012e-05 4.36214035e-05 4.06814669e-05\n",
            " 3.79396722e-05 3.53826653e-05 3.29979921e-05 3.07740378e-05\n",
            " 2.86999707e-05 2.67656887e-05 2.49617709e-05 2.32794311e-05\n",
            " 2.17104753e-05 2.02472620e-05 1.88826643e-05 1.76100359e-05] \n",
            "\n",
            "Training Error of Elastic Regression Model : 0.9948646996252336 \n",
            "\n",
            "Generalisation Error of Elastic Regression Model : 1.5502852179057978\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def elastic(X_train, y_train, X_test, y_test, alpha):\n",
        "    elastic = ElasticNet(alpha = alpha, l1_ratio = 0.5, normalize = True)\n",
        "    elastic.fit(X_train, y_train)\n",
        "    return elastic.predict(X_test)\n",
        "\n",
        "alphas = np.array(elastic_regression_cv.alphas_)\n",
        "\n",
        "x = np.array(alphas)\n",
        "x_log = np.array([math.log(alpha, 10) for alpha in alphas])\n",
        "y = np.array([metrics.mean_squared_error(y_test, elastic(X_train, y_train, X_test, y_test, alpha)) for alpha in alphas])\n",
        "\n",
        "plt.plot(x, y)\n",
        "plt.title(\"Generalisation Error v/s Regularisation parameter\")\n",
        "plt.xlabel(\"Regularisation Parameter\")\n",
        "plt.ylabel(\"Generalisation Error\")\n",
        "plt.show()\n",
        "\n",
        "plt.plot(x_log, y)\n",
        "plt.title(\"Generalisation Error v/s Logarithm of Regularisation parameter\")\n",
        "plt.xlabel(\"log(Regularisation Parameter)\")\n",
        "plt.ylabel(\"Generalisation Error\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "_bPahcWLCb9Q",
        "outputId": "b94222a5-43d5-445e-c8af-6406e997d49f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}