week3_auroc.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.metrics import roc_curve, auc\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.metrics import roc_curve, auc\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "\n",
        "n_samples = 300\n",
        "\n",
        "fig, axes = plt.subplots(1, 2, figsize=(8, 4))\n",
        "labels = np.random.choice([0, 1], size=n_samples, p=[0.8, 0.2])\n",
        "predictions = labels + 0.5 * np.random.randn(n_samples)\n",
        "\n",
        "plt.sca(axes[0])\n",
        "plt.hist(predictions[labels == 0], alpha = 0.5)\n",
        "plt.hist(predictions[labels == 1], alpha = 0.5)\n",
        "plt.xlabel('logits')\n",
        "\n",
        "plt.sca(axes[1])\n",
        "df_plot = pd.DataFrame({'x': predictions, 'y': labels})\n",
        "sns.boxplot(data = df_plot, x = 'y', y = 'x')\n",
        "plt.plot()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 406
        },
        "id": "iZEOghpCK76v",
        "outputId": "87374ddf-4233-4dd4-f1a3-a7a5b9bcbbe4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[]"
            ]
          },
          "metadata": {},
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x400 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J72KqR4NVcrU",
        "outputId": "6d6a10c9-6996-40c6-9a1b-fc47a13de5e7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0,\n",
              "       0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0,\n",
              "       0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0,\n",
              "       1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0,\n",
              "       1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0,\n",
              "       0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1,\n",
              "       0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0,\n",
              "       0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0,\n",
              "       1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1])"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QFXwCWMeVeOy",
        "outputId": "39213405-a889-4a68-bdb6-d67ff73e3377"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 1.14690407e-01,  3.69601685e-01,  6.69191719e-01,  1.36069015e+00,\n",
              "       -1.21178456e-01,  1.93658455e-01, -1.33688781e+00, -9.47008191e-02,\n",
              "        1.85818747e+00,  7.46397657e-01,  1.23481111e+00,  1.62568697e+00,\n",
              "        2.12763960e-01,  7.90069077e-01,  1.24233603e+00,  2.36312358e-01,\n",
              "       -1.63729408e+00, -5.06606229e-01, -1.01577476e+00, -3.58010341e-01,\n",
              "        4.65336982e-01,  1.37085090e+00,  5.58979390e-02, -5.37332499e-01,\n",
              "       -8.37105920e-01,  1.50876850e+00,  7.98085770e-01,  1.04013492e+00,\n",
              "       -1.77552033e+00, -5.83172338e-01,  2.44207081e-01, -7.40671226e-01,\n",
              "       -9.23276308e-01, -4.55600923e-01,  1.17687031e+00,  2.01270843e+00,\n",
              "       -3.75787954e-01, -7.82633652e-02,  8.43920968e-01,  5.66753438e-01,\n",
              "        1.99592812e+00,  4.72540378e-01, -7.24121045e-01,  1.53337188e-01,\n",
              "       -1.16079198e+00,  7.98593744e-01, -2.53527759e-01, -4.02717682e-01,\n",
              "        6.13880668e-01, -1.12578643e-01,  2.77314764e-02,  7.62554126e-01,\n",
              "        9.43514027e-01,  3.23635169e-01,  1.22830507e-01, -6.31116212e-01,\n",
              "       -2.93195673e-01,  7.14635042e-01, -3.20788370e-01,  3.45427029e-01,\n",
              "       -1.80435275e-01,  1.87920635e-03,  9.09220297e-01,  1.21470414e+00,\n",
              "        4.14047611e-01,  4.28343217e-01,  8.43653267e-01,  6.84900673e-01,\n",
              "        2.53939218e+00,  2.29490084e-01, -1.16148566e+00,  9.57622908e-01,\n",
              "        1.48983405e-01, -1.77569080e+00, -5.88735045e-01, -2.69783464e-01,\n",
              "        9.10499323e-01,  4.52471875e-01,  4.75036910e-01, -1.33717454e+00,\n",
              "        9.16028038e-01,  1.52899074e+00,  1.89097189e+00, -4.80842452e-01,\n",
              "       -7.18382548e-01, -3.55049355e-02,  3.68343453e-01, -1.95492675e-03,\n",
              "        1.89118222e+00,  1.39953772e+00,  1.55076076e+00, -6.07207700e-01,\n",
              "        1.67739406e+00,  5.22656275e-01,  3.12668407e-01, -5.96942058e-01,\n",
              "        4.32881006e-01,  3.95888831e-01,  7.77532973e-01,  1.35413957e+00,\n",
              "        1.58686429e+00,  1.78433676e+00,  7.53910988e-01, -5.74999439e-01,\n",
              "       -1.96938644e-01,  3.14706786e-01,  3.97612813e-01, -1.30320524e-01,\n",
              "       -6.63844445e-01,  9.45299892e-01, -5.09719480e-01,  3.48554030e-03,\n",
              "       -1.55712804e-01, -2.17120881e+00, -2.05986640e-01,  8.19063236e-01,\n",
              "       -9.35645508e-01, -2.93972017e-01,  3.56344663e-01, -1.38838557e+00,\n",
              "        2.09270310e-01, -2.63420290e-01,  5.25475564e-01,  1.34620086e+00,\n",
              "        6.29018417e-01,  4.12397936e-01, -5.35172893e-01,  7.66218812e-01,\n",
              "       -1.31482951e+00,  5.20569779e-02, -1.43853506e-01, -7.27787843e-01,\n",
              "       -1.87910956e-02, -3.32388728e-01,  6.28591817e-01, -6.14875236e-01,\n",
              "        5.01934515e-01,  1.04848038e+00, -8.13385186e-01,  6.84167145e-01,\n",
              "        7.07154267e-01, -1.22141574e+00, -3.02211836e-01,  7.82238629e-01,\n",
              "       -3.27001007e-01,  2.40846283e-01,  5.80681164e-01,  1.23587953e+00,\n",
              "        1.00831911e+00, -8.71251121e-01, -3.31966850e-01,  3.85990385e-02,\n",
              "        3.49164993e-02, -8.64972751e-01,  1.77471053e+00, -1.96291799e-01,\n",
              "        8.74916441e-02, -4.13332313e-01,  3.44626045e-01,  2.15786144e+00,\n",
              "        8.07375713e-01,  3.28327519e-01, -1.14727967e+00, -1.55961662e+00,\n",
              "       -6.73269661e-01,  1.36224271e+00,  1.17432418e+00,  4.16020103e-01,\n",
              "        1.06146002e+00,  8.18894334e-01, -1.02613879e+00,  2.15429472e+00,\n",
              "        2.88196455e-01, -8.18829115e-02,  1.25664013e+00, -1.30353219e+00,\n",
              "       -6.70628134e-01,  2.07859557e+00,  4.15376521e-01,  3.77186657e-01,\n",
              "       -1.44417482e-01,  1.24668864e-01, -1.19395668e+00, -2.70378594e-01,\n",
              "        5.28377373e-01, -1.38746275e-01,  1.35098203e+00,  1.47452526e-01,\n",
              "       -1.12974306e-01,  3.11670574e-01,  1.78881753e+00, -1.07340619e-01,\n",
              "        2.59219893e-01,  1.93892673e-01,  8.22803065e-01,  2.61421837e-01,\n",
              "        2.16267474e+00, -8.49093899e-01, -1.41527403e-01, -3.96156193e-01,\n",
              "        1.32615461e-01, -3.29183194e-03, -3.49021538e-01, -1.60986390e+00,\n",
              "        1.66089226e+00,  1.30185815e+00, -1.15010163e-01,  3.71332473e-01,\n",
              "        8.71067405e-01,  8.21365822e-01,  1.22239145e+00,  1.20188127e-01,\n",
              "        2.45259098e-01, -1.35842005e+00, -8.99683255e-01,  2.76172444e+00,\n",
              "       -7.76953754e-01,  1.23285181e+00,  1.38027376e-01,  1.04829476e+00,\n",
              "        3.95821100e-02,  1.90988592e-01, -1.09419055e+00,  1.62085622e+00,\n",
              "        1.31394593e-01,  2.58045119e-01,  1.59300968e+00, -1.27282810e+00,\n",
              "        8.05752866e-01,  2.18459320e+00,  1.60155536e+00,  1.46488023e+00,\n",
              "       -2.26336547e+00,  6.01874857e-02,  4.31702903e-01, -8.14283017e-01,\n",
              "       -5.91271853e-01,  1.96748000e+00,  1.67950258e-01,  4.37243586e-01,\n",
              "       -6.63551028e-01,  1.02340958e+00, -6.78574960e-01, -1.44761092e-01,\n",
              "        1.27807418e+00,  1.18833665e+00,  2.95165376e-01,  7.54105990e-01,\n",
              "        1.36754280e-02, -5.77384639e-01, -9.57951374e-03,  3.52048973e-01,\n",
              "       -1.10241316e+00,  2.11117381e+00, -1.94614699e-01, -5.70031403e-01,\n",
              "        1.50093968e+00,  1.68200502e-01, -4.71488910e-01,  1.06372311e+00,\n",
              "        8.82857062e-01,  6.79676143e-01,  1.58187094e+00, -1.99956416e+00,\n",
              "        1.45269930e+00,  1.05362954e+00, -7.48841242e-01, -1.69909597e+00,\n",
              "        1.45767726e+00, -1.62175300e+00,  3.36740647e-01, -1.24738307e-01,\n",
              "       -1.62354317e+00,  3.09532068e-01,  2.05460303e+00,  1.99126691e+00,\n",
              "       -8.66289782e-01,  2.08975450e+00,  1.15603633e+00, -1.65836068e+00,\n",
              "        3.69263053e-01,  6.77370755e-01, -5.49239645e-01,  9.85535091e-01,\n",
              "       -3.67051567e-01,  1.59163409e+00, -5.63053059e-02,  1.76027336e-01,\n",
              "        3.24510848e-01,  6.99845050e-01,  7.96413667e-01, -9.82335127e-01,\n",
              "       -5.44624200e-01, -9.48688368e-02,  2.39995540e-01,  4.24393667e-01,\n",
              "        7.12691312e-01,  8.88561394e-01, -2.58519203e-01,  1.21624563e+00])"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "c = 0.5\n",
        "acc = np.mean((predictions > c) == labels)\n",
        "print(f'cutoff: {c}, accuracy: {acc}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V4fJm7F7MbgQ",
        "outputId": "2b71439c-8d55-424b-babc-5daea04fdba7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "cutoff: 0.5, accuracy: 0.83\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "c = 0.8\n",
        "acc = np.mean((predictions > c) == labels)\n",
        "print(f'cutoff: {c}, accuracy: {acc}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4f9tZH_oMQn0",
        "outputId": "234aa2fe-a028-4e12-8e53-b176dae43805"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "cutoff: 0.8, accuracy: 0.8866666666666667\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "c = 0.2\n",
        "acc = np.mean((predictions > c) == labels)\n",
        "print(f'cutoff: {c}, accuracy: {acc}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C3lpwzAMMQ5e",
        "outputId": "0d89fc5f-7149-4396-9e0b-a31b89303e3b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "cutoff: 0.2, accuracy: 0.7233333333333334\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "res = []\n",
        "for c in [0.1, 0.25, 0.5, 0.75, 0.9, 0.99]:\n",
        "  pred = (predictions > c)\n",
        "  acc = np.mean(pred == labels)\n",
        "  precision = np.mean(labels[pred])\n",
        "  TPR = np.mean(pred[labels == 1])\n",
        "  FNR = np.mean(pred[labels == 0])\n",
        "  res.append([c, acc, precision, TPR, FNR])\n",
        "pd.DataFrame(res, columns = ['cutoff', 'accuracy', 'precision', 'TPR/recall', 'FNR'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "kXuXocdeM3Lr",
        "outputId": "b9278631-2654-4178-c77b-c72514b6d4c0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   cutoff  accuracy  precision  TPR/recall       FNR\n",
              "0    0.10  0.693333   0.421384    1.000000  0.394850\n",
              "1    0.25  0.746667   0.467153    0.955224  0.313305\n",
              "2    0.50  0.830000   0.576923    0.895522  0.188841\n",
              "3    0.75  0.873333   0.716418    0.716418  0.081545\n",
              "4    0.90  0.886667   0.811321    0.641791  0.042918\n",
              "5    0.99  0.896667   0.875000    0.626866  0.025751"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-0cd4515c-9ecf-43a8-accb-2d128705d754\" class=\"colab-df-container\">\n",
              "    <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>cutoff</th>\n",
              "      <th>accuracy</th>\n",
              "      <th>precision</th>\n",
              "      <th>TPR/recall</th>\n",
              "      <th>FNR</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.10</td>\n",
              "      <td>0.693333</td>\n",
              "      <td>0.421384</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.394850</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.25</td>\n",
              "      <td>0.746667</td>\n",
              "      <td>0.467153</td>\n",
              "      <td>0.955224</td>\n",
              "      <td>0.313305</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.50</td>\n",
              "      <td>0.830000</td>\n",
              "      <td>0.576923</td>\n",
              "      <td>0.895522</td>\n",
              "      <td>0.188841</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.75</td>\n",
              "      <td>0.873333</td>\n",
              "      <td>0.716418</td>\n",
              "      <td>0.716418</td>\n",
              "      <td>0.081545</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.90</td>\n",
              "      <td>0.886667</td>\n",
              "      <td>0.811321</td>\n",
              "      <td>0.641791</td>\n",
              "      <td>0.042918</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>0.99</td>\n",
              "      <td>0.896667</td>\n",
              "      <td>0.875000</td>\n",
              "      <td>0.626866</td>\n",
              "      <td>0.025751</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-0cd4515c-9ecf-43a8-accb-2d128705d754')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-0cd4515c-9ecf-43a8-accb-2d128705d754 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-0cd4515c-9ecf-43a8-accb-2d128705d754');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-284be32b-d80b-431e-8213-236ece9b11fb\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-284be32b-d80b-431e-8213-236ece9b11fb')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-284be32b-d80b-431e-8213-236ece9b11fb button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"pd\",\n  \"rows\": 6,\n  \"fields\": [\n    {\n      \"column\": \"cutoff\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.35918890109059143,\n        \"min\": 0.1,\n        \"max\": 0.99,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          0.1,\n          0.25,\n          0.99\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"accuracy\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.08328443009528427,\n        \"min\": 0.6933333333333334,\n        \"max\": 0.8966666666666666,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          0.6933333333333334,\n          0.7466666666666667,\n          0.8966666666666666\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"precision\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.1854350372326308,\n        \"min\": 0.42138364779874216,\n        \"max\": 0.875,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          0.42138364779874216,\n          0.46715328467153283,\n          0.875\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"TPR/recall\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.1643147336659449,\n        \"min\": 0.6268656716417911,\n        \"max\": 1.0,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          1.0,\n          0.9552238805970149,\n          0.6268656716417911\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"FNR\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.15238569982666644,\n        \"min\": 0.02575107296137339,\n        \"max\": 0.3948497854077253,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          0.3948497854077253,\n          0.3133047210300429,\n          0.02575107296137339\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.metrics import roc_curve, auc\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.metrics import roc_curve, auc\n",
        "import pandas as pd\n",
        "\n",
        "# Function to plot ROC curve and calculate AUROC\n",
        "def plot_roc_curve(labels, predictions, title='ROC Curve'):\n",
        "    fpr, tpr, _ = roc_curve(labels, predictions)\n",
        "    roc_auc = auc(fpr, tpr)\n",
        "    #plt.figure()\n",
        "    plt.plot(fpr, tpr, color='darkorange', lw=2, label=f'AUROC = {roc_auc:.2f}')\n",
        "    plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')\n",
        "    plt.xlim([0.0, 1.0])\n",
        "    plt.ylim([0.0, 1.0])\n",
        "    plt.xlabel('False Positive Rate')\n",
        "    plt.ylabel('True Positive Rate')\n",
        "    plt.title(title)\n",
        "    plt.legend(loc='lower right')\n",
        "    #plt.show()\n",
        "\n",
        "n_samples = 300\n",
        "\n",
        "plot_roc_curve(labels, predictions, title='aucroc')\n",
        "\n",
        "res = []\n",
        "for c in [0.1, 0.5, 0.9]:\n",
        "  TPR = np.mean((predictions > c)[labels == 1])\n",
        "  FNR = np.mean((predictions > c)[labels == 0])\n",
        "  plt.plot(FNR, TPR, '.', markersize=15)\n",
        "  res.append([c, FNR, TPR])\n",
        "print(pd.DataFrame(res, columns = ['cutoff', 'x:FNR', 'y:TPR']))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 541
        },
        "id": "dlqhy-AiK7kX",
        "outputId": "32a27a67-af36-4929-915d-4b70dda1a722"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "   cutoff     x:FNR     y:TPR\n",
            "0     0.1  0.394850  1.000000\n",
            "1     0.5  0.188841  0.895522\n",
            "2     0.9  0.042918  0.641791\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "n_samples = 300\n",
        "\n",
        "fig, axes = plt.subplots(2, 3, figsize=(10, 6))\n",
        "\n",
        "for ip, p in enumerate([0.1, 0.5, 0.9]):\n",
        "  labels = np.random.choice([0, 1], size=n_samples, p=[1-p, p])\n",
        "  predictions = labels + 0.9 * np.random.randn(n_samples)\n",
        "\n",
        "  plt.sca(axes[1, ip])\n",
        "  plt.hist(predictions[labels == 0])\n",
        "  plt.hist(predictions[labels == 1])\n",
        "  plt.sca(axes[0, ip])\n",
        "  plot_roc_curve(labels, predictions, title=f'%.3f Positives' % p)\n",
        "  print(np.sum((predictions > 0) == labels) / n_samples)\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 659
        },
        "id": "MpS5aTg9zGJf",
        "outputId": "1af34d22-85ab-4762-d052-c20d6135080a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.55\n",
            "0.74\n",
            "0.8333333333333334\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 6 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import roc_auc_score\n",
        "import pandas as pd\n",
        "n_samples = 300\n",
        "\n",
        "table = []\n",
        "for ip, p in enumerate([0.1, 0.5, 0.9]):\n",
        "  labels = np.random.choice([0, 1], size=n_samples, p=[1-p, p])\n",
        "  pred_ = labels + 0.7 * np.random.randn(n_samples)\n",
        "  pred_cutoff = pred_ > 0.5\n",
        "  pred_allzero = np.zeros(n_samples)\n",
        "  pred_random = np.random.choice([0, 1], size=n_samples, p=[0.5, 0.5])\n",
        "\n",
        "  acc_cutoff = np.sum(pred_cutoff == labels) / n_samples\n",
        "  acc_allzero = np.sum(pred_allzero == labels) / n_samples\n",
        "  acc_random = np.sum(pred_random == labels) / n_samples\n",
        "\n",
        "  auroc_cutoff = roc_auc_score(labels, pred_)\n",
        "  auroc_allzero = roc_auc_score(labels, pred_allzero)\n",
        "  auroc_random = roc_auc_score(labels, pred_random)\n",
        "\n",
        "  table += [[p, acc_cutoff, acc_allzero, acc_random, auroc_cutoff, auroc_allzero, auroc_random]]\n",
        "pd.DataFrame(table, columns = ['%positives', 'acc_cutoff', 'acc_allzero', 'acc_random', 'aucroc_cutoff', 'aucroc_allzero', 'aucroc_random'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 143
        },
        "id": "ZMn8J8yi3v0Q",
        "outputId": "a108edd5-3b8b-43c9-d478-a3b786241100"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   %positives  acc_cutoff  acc_allzero  acc_random  aucroc_cutoff  \\\n",
              "0         0.1    0.763333     0.890000    0.530000       0.865850   \n",
              "1         0.5    0.783333     0.453333    0.503333       0.853434   \n",
              "2         0.9    0.736667     0.113333    0.503333       0.835029   \n",
              "\n",
              "   aucroc_allzero  aucroc_random  \n",
              "0             0.5       0.457099  \n",
              "1             0.5       0.506815  \n",
              "2             0.5       0.553184  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-cf39f7e3-de01-4bd0-ae58-9e2cad914727\" class=\"colab-df-container\">\n",
              "    <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>%positives</th>\n",
              "      <th>acc_cutoff</th>\n",
              "      <th>acc_allzero</th>\n",
              "      <th>acc_random</th>\n",
              "      <th>aucroc_cutoff</th>\n",
              "      <th>aucroc_allzero</th>\n",
              "      <th>aucroc_random</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.1</td>\n",
              "      <td>0.763333</td>\n",
              "      <td>0.890000</td>\n",
              "      <td>0.530000</td>\n",
              "      <td>0.865850</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.457099</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.5</td>\n",
              "      <td>0.783333</td>\n",
              "      <td>0.453333</td>\n",
              "      <td>0.503333</td>\n",
              "      <td>0.853434</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.506815</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.9</td>\n",
              "      <td>0.736667</td>\n",
              "      <td>0.113333</td>\n",
              "      <td>0.503333</td>\n",
              "      <td>0.835029</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.553184</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-cf39f7e3-de01-4bd0-ae58-9e2cad914727')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-cf39f7e3-de01-4bd0-ae58-9e2cad914727 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-cf39f7e3-de01-4bd0-ae58-9e2cad914727');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-a0639a40-e36c-482c-9ee3-9904abd6b584\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-a0639a40-e36c-482c-9ee3-9904abd6b584')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-a0639a40-e36c-482c-9ee3-9904abd6b584 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"pd\",\n  \"rows\": 3,\n  \"fields\": [\n    {\n      \"column\": \"%positives\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.4,\n        \"min\": 0.1,\n        \"max\": 0.9,\n        \"num_unique_values\": 3,\n        \"samples\": [\n          0.1,\n          0.5,\n          0.9\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"acc_cutoff\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.02341256389522829,\n        \"min\": 0.7366666666666667,\n        \"max\": 0.7833333333333333,\n        \"num_unique_values\": 3,\n        \"samples\": [\n          0.7633333333333333,\n          0.7833333333333333,\n          0.7366666666666667\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"acc_allzero\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.3893346651423188,\n        \"min\": 0.11333333333333333,\n        \"max\": 0.89,\n        \"num_unique_values\": 3,\n        \"samples\": [\n          0.89,\n          0.4533333333333333,\n          0.11333333333333333\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"acc_random\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.015396007178390054,\n        \"min\": 0.5033333333333333,\n        \"max\": 0.53,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          0.5033333333333333,\n          0.53\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"aucroc_cutoff\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.01550710341418429,\n        \"min\": 0.8350287483414419,\n        \"max\": 0.8658495062989445,\n        \"num_unique_values\": 3,\n        \"samples\": [\n          0.8658495062989445,\n          0.8534343615494978\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"aucroc_allzero\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.0,\n        \"min\": 0.5,\n        \"max\": 0.5,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          0.5\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"aucroc_random\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.048052386290003524,\n        \"min\": 0.4570990806945864,\n        \"max\": 0.5531844316674038,\n        \"num_unique_values\": 3,\n        \"samples\": [\n          0.4570990806945864\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "table = []\n",
        "n_samples = 1000\n",
        "labels = np.random.choice([0, 1], size=n_samples, p=[0.99, 0.01])\n",
        "for ip, p in enumerate([0.3, 0.5, 0.8, 0.9]):\n",
        "  pred_ = labels + p * np.random.randn(n_samples)\n",
        "  pred_cutoff = pred_ > 0.5\n",
        "\n",
        "  acc = np.sum(pred_cutoff == labels) / n_samples\n",
        "  precision = np.sum((pred_cutoff == 1) & (labels == 1)) / np.sum(pred_cutoff == 1)\n",
        "  recall = np.sum((pred_cutoff == 1) & (labels == 1)) / np.sum(labels == 1)\n",
        "  auroc = roc_auc_score(labels, pred_)\n",
        "\n",
        "  table += [[p, acc, precision, recall, auroc]]\n",
        "pd.DataFrame(table, columns = ['noise level', 'acc', 'precision', 'recall', 'aucroc'])"
      ],
      "metadata": {
        "id": "fcBFjVq0dsc8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 174
        },
        "outputId": "3e7c79ce-67f0-469c-f3b4-ebc7dd43a682"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   noise level    acc  precision    recall    aucroc\n",
              "0          0.3  0.951   0.172414  0.909091  0.974998\n",
              "1          0.5  0.849   0.056962  0.818182  0.897325\n",
              "2          0.8  0.751   0.035156  0.818182  0.890891\n",
              "3          0.9  0.723   0.028369  0.727273  0.792352"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-197581a1-540f-429b-860b-d1bd332d3a9a\" class=\"colab-df-container\">\n",
              "    <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>noise level</th>\n",
              "      <th>acc</th>\n",
              "      <th>precision</th>\n",
              "      <th>recall</th>\n",
              "      <th>aucroc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.3</td>\n",
              "      <td>0.951</td>\n",
              "      <td>0.172414</td>\n",
              "      <td>0.909091</td>\n",
              "      <td>0.974998</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.5</td>\n",
              "      <td>0.849</td>\n",
              "      <td>0.056962</td>\n",
              "      <td>0.818182</td>\n",
              "      <td>0.897325</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.8</td>\n",
              "      <td>0.751</td>\n",
              "      <td>0.035156</td>\n",
              "      <td>0.818182</td>\n",
              "      <td>0.890891</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.9</td>\n",
              "      <td>0.723</td>\n",
              "      <td>0.028369</td>\n",
              "      <td>0.727273</td>\n",
              "      <td>0.792352</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-197581a1-540f-429b-860b-d1bd332d3a9a')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-197581a1-540f-429b-860b-d1bd332d3a9a button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-197581a1-540f-429b-860b-d1bd332d3a9a');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-1063db02-dd45-444a-b350-b5a8a390aae1\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-1063db02-dd45-444a-b350-b5a8a390aae1')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-1063db02-dd45-444a-b350-b5a8a390aae1 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"pd\",\n  \"rows\": 4,\n  \"fields\": [\n    {\n      \"column\": \"noise level\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.2753785273643051,\n        \"min\": 0.3,\n        \"max\": 0.9,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          0.5,\n          0.9,\n          0.3\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"acc\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.10354226190305095,\n        \"min\": 0.723,\n        \"max\": 0.951,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          0.849,\n          0.723,\n          0.951\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"precision\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.06724138095126454,\n        \"min\": 0.028368794326241134,\n        \"max\": 0.1724137931034483,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          0.056962025316455694,\n          0.028368794326241134,\n          0.1724137931034483\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"recall\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.07422696190252052,\n        \"min\": 0.7272727272727273,\n        \"max\": 0.9090909090909091,\n        \"num_unique_values\": 3,\n        \"samples\": [\n          0.9090909090909091,\n          0.8181818181818182,\n          0.7272727272727273\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"aucroc\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.07485370149930992,\n        \"min\": 0.7923522382571928,\n        \"max\": 0.9749977019946686,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          0.8973251217942826,\n          0.7923522382571928,\n          0.9749977019946686\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "0vZAS1An1_-i"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}