Project_5.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Project 5.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPyuzp4eGMOqDKHZY253fpy",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/rebecca-yoo/github-slideshow/blob/main/Project_5.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-8G_dUDdOejp"
      },
      "source": [
        "# Project 5\n",
        "\n",
        "1. Fit a pytorch model to the oasis data that treats gold lesions as the outcome and T2 value as the predictor. Use a bias term, sigmoid output activation function and no hidden layers. Do not split the data into a testing and training dataset. Demonstrate that the fitted values are the same as those obtained with logistic regression.  \n",
        "\n",
        "2. Fit a pytorch neural network with two hidden layers on the diamond dataset from the previous project predicting price. Evaluate your predictions on held out test set data."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 720
        },
        "id": "1p1BZfbwOdyB",
        "outputId": "ca15c735-b898-4dfd-d877-7a5f909a8a02"
      },
      "source": [
        "import pandas as pd\n",
        "import torch\n",
        "import statsmodels.formula.api as smf\n",
        "import statsmodels.api as sm\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "## this sets some style parameters\n",
        "sns.set()\n",
        "\n",
        "## Download in the data if it's not already there\n",
        "! if [ ! -e oasis.csv ]; \\\n",
        "  then wget https://raw.githubusercontent.com/bcaffo/ds4bme_intro/master/data/oasis.csv; \\\n",
        "fi;\n",
        "\n",
        "dat = pd.read_csv('https://raw.githubusercontent.com/bcaffo/ds4bme_intro/master/data/oasis.csv')\n",
        "\n",
        "## Create a normalized regression variable\n",
        "dat = dat.assign(y = dat.GOLD_Lesions)\n",
        "dat = dat.assign(x = (dat.T2 - np.mean(dat.T2)) / np.std(dat.T2))\n",
        "\n",
        "fit = smf.logit('y~x', data=dat).fit()\n",
        "print(fit.summary())\n",
        "\n",
        "# The in sample predictions\n",
        "yhat = 1 / (1 + np.exp(-fit.fittedvalues))\n",
        "\n",
        "n = dat.shape[0]\n",
        "\n",
        "## Get the y and x from \n",
        "xtraining = torch.from_numpy(dat['x'].values)\n",
        "ytraining = torch.from_numpy(dat['y'].values)\n",
        "\n",
        "## PT wants floats\n",
        "xtraining = xtraining.float()\n",
        "ytraining = ytraining.float()\n",
        "\n",
        "## Dimension is 1xn not nx1\n",
        "## squeeze the second dimension\n",
        "xtraining = xtraining.unsqueeze(1)\n",
        "ytraining = ytraining.unsqueeze(1)\n",
        "\n",
        "## Show that everything is the right size\n",
        "print([xtraining.shape, \n",
        " ytraining.shape,\n",
        " [n, 1]\n",
        " ])\n",
        "\n",
        "class LogisticRegression(torch.nn.Module):\n",
        "     def __init__(self):\n",
        "        super(LogisticRegression, self).__init__()\n",
        "        self.linear = torch.nn.Linear(1, 1, bias = True)\n",
        "     def forward(self, x):\n",
        "        y_pred = torch.sigmoid(self.linear(x))\n",
        "        return y_pred\n",
        "\n",
        "## Then the model is simply  \n",
        "model = LogisticRegression()\n",
        "\n",
        "## MSE is the loss function\n",
        "loss_fn = torch.nn.BCELoss()  \n",
        "\n",
        "## Set the optimizer\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=1e-2)\n",
        "\n",
        "## Loop over iterations\n",
        "for t in range(10000):\n",
        "\n",
        "  ## Forward propagation\n",
        "  y_pred = model(xtraining)\n",
        "\n",
        "  ## the loss for this interation\n",
        "  loss = loss_fn(y_pred, ytraining)\n",
        "\n",
        "  ## Zero out the gradients before adding them up \n",
        "  optimizer.zero_grad()\n",
        "  \n",
        "  ## Backprop\n",
        "  loss.backward()\n",
        "  \n",
        "  ## Optimization step\n",
        "  optimizer.step()\n",
        "\n",
        "ytest = model(xtraining)\n",
        "ytest = ytest.detach().numpy().reshape(-1)\n",
        "plt.plot(yhat, ytest, '.')\n",
        "plt.plot([0.4, 0.7], [0.4, 0.7], linewidth=2)\n",
        "plt.title('Logistic Regression Predictions (yhat) vs. Sigmoid Pytorch Model Predictions (ytest)')\n",
        "plt.xlabel('Yhat')\n",
        "plt.ylabel('Ytest')\n",
        "for param in model.parameters():  \n",
        "  print(param.data)\n"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
            "  import pandas.util.testing as tm\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Optimization terminated successfully.\n",
            "         Current function value: 0.687421\n",
            "         Iterations 4\n",
            "                           Logit Regression Results                           \n",
            "==============================================================================\n",
            "Dep. Variable:                      y   No. Observations:                  100\n",
            "Model:                          Logit   Df Residuals:                       98\n",
            "Method:                           MLE   Df Model:                            1\n",
            "Date:                Mon, 04 Oct 2021   Pseudo R-squ.:                0.008262\n",
            "Time:                        20:51:44   Log-Likelihood:                -68.742\n",
            "converged:                       True   LL-Null:                       -69.315\n",
            "Covariance Type:            nonrobust   LLR p-value:                    0.2845\n",
            "==============================================================================\n",
            "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
            "------------------------------------------------------------------------------\n",
            "Intercept      0.0005      0.201      0.002      0.998      -0.394       0.395\n",
            "x              0.2159      0.204      1.061      0.289      -0.183       0.615\n",
            "==============================================================================\n",
            "[torch.Size([100, 1]), torch.Size([100, 1]), [100, 1]]\n",
            "tensor([[0.2159]])\n",
            "tensor([0.0005])\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 728
        },
        "id": "DPYpklvbtemw",
        "outputId": "1a2f13ec-0c02-49b9-e406-7283098135a6"
      },
      "source": [
        "N, H1, H2 = 53940, 18, 18\n",
        "\n",
        "df = pd.read_csv('https://raw.githubusercontent.com/tidyverse/ggplot2/master/data-raw/diamonds.csv')\n",
        "\n",
        "# transform data\n",
        "\n",
        "df = pd.get_dummies(df, columns=[\"cut\"], prefix=\"Cut\", drop_first = True)\n",
        "df = pd.get_dummies(df, columns=[\"color\"], prefix=\"Color\", drop_first = True)\n",
        "df = pd.get_dummies(df, columns=[\"clarity\"], prefix=\"Clarity\", drop_first = True)\n",
        "\n",
        "y_1 = df['price']\n",
        "x_1 = df.drop(['depth', 'table', 'price', 'x', 'y', 'z'], axis = 1)\n",
        "\n",
        "y_1 = torch.from_numpy(y_1.values)\n",
        "x_1 = torch.from_numpy(x_1.values)\n",
        "\n",
        "trainFraction = 0.75\n",
        "\n",
        "## Hold out data\n",
        "sample = np.random.uniform(size = 53940) < trainFraction\n",
        "xtrain_1 = x_1[sample]\n",
        "ytrain_1 = y_1[sample]\n",
        "xtest_1 =  x_1[~sample]\n",
        "ytest_1 =  y_1[~sample]\n",
        "\n",
        "## PT wants floats\n",
        "xtrain_1 = xtrain_1.float()\n",
        "ytrain_1 = ytrain_1.float()\n",
        "ytrain_1_val = ytrain_1\n",
        "xtest_1 = xtest_1.float()\n",
        "ytest_1 = ytest_1.float()\n",
        "\n",
        "## Dimension is 1xn not nx1\n",
        "## squeeze the second dimension\n",
        "ytrain_1 = ytrain_1.unsqueeze(1)\n",
        "ytest_1 = ytest_1.unsqueeze(1)\n",
        "\n",
        "print([\n",
        " xtrain_1.shape,\n",
        " ytrain_1.shape,\n",
        " xtest_1.shape,\n",
        " ytest_1.shape,\n",
        "])\n",
        "\n",
        "## Define the model\n",
        "## Dimension of the hidden layer\n",
        "\n",
        "## Number of predictors\n",
        "D_in = xtrain_1.shape[1]\n",
        "D_out = 1\n",
        "model_1 = torch.nn.Sequential(\n",
        "    torch.nn.Linear(D_in, H1),\n",
        "    torch.nn.ReLU(),\n",
        "    torch.nn.Linear(H1, H2),\n",
        "    torch.nn.ReLU(),\n",
        "    torch.nn.Linear(H2, D_out)\n",
        ")\n",
        "\n",
        "loss_fn_1 = torch.nn.MSELoss(reduction='mean')\n",
        "\n",
        "optimizer_1 = torch.optim.Adam(model.parameters(), lr=1e-5)\n",
        "\n",
        "for j in range(500):\n",
        "  for k in range(xtrain_1.shape[0]//128):\n",
        "    ytrain_pred_1 = model_1(xtrain_1[k*128:k*128+128])\n",
        "    loss_1 = loss_fn_1(ytrain_pred_1, ytrain_1[k*128:k*128+128])\n",
        "    #if j % 100 == 99:\n",
        "      # print(j, loss.item())\n",
        "    optimizer_1.zero_grad()\n",
        "    loss_1.backward()\n",
        "    optimizer_1.step()\n",
        "\n",
        "## try prediction\n",
        "ytesting_pred_1 = model_1(xtest_1)\n",
        "a = ytesting_pred_1.detach().numpy().reshape(-1)\n",
        "\n",
        "plt.scatter(a, ytest_1[:,0])\n",
        "plt.title('Predicted Outcomes of Price (y_hat) vs. Ground Truth Values of Price (y_test)')\n",
        "plt.xlabel('Predicted Outcomes')\n",
        "plt.ylabel('Ground Truth Values')\n"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[torch.Size([40446, 18]), torch.Size([40446, 1]), torch.Size([13494, 18]), torch.Size([13494, 1])]\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "TypeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-6-dca73a900f58>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     80\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mylabel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Ground Truth Values'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     81\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 82\u001b[0;31m \u001b[0mmse\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msqrt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmean\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mytrain_1_val\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mytrain_pred_1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     83\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'The RMSE is '\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<__array_function__ internals>\u001b[0m in \u001b[0;36mmean\u001b[0;34m(*args, **kwargs)\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/numpy/core/fromnumeric.py\u001b[0m in \u001b[0;36mmean\u001b[0;34m(a, axis, dtype, out, keepdims)\u001b[0m\n\u001b[1;32m   3368\u001b[0m             \u001b[0;32mpass\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3369\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 3370\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mmean\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mout\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mout\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   3371\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3372\u001b[0m     return _methods._mean(a, axis=axis, dtype=dtype,\n",
            "\u001b[0;31mTypeError\u001b[0m: mean() received an invalid combination of arguments - got (out=NoneType, axis=NoneType, dtype=NoneType, ), but expected one of:\n * (*, torch.dtype dtype)\n * (tuple of ints dim, bool keepdim, *, torch.dtype dtype)\n * (tuple of names dim, bool keepdim, *, torch.dtype dtype)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    }
  ]
}