logistic_regression.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Copy of logistic_regression.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [
        "JndnmDMp66FL",
        "dPpJUV862FYI",
        "i2e3TlyL57Qs",
        "wCugvl0JdWYL"
      ],
      "include_colab_link": true
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "[View in Colaboratory](https://colab.research.google.com/github/DillipKS/MLCC_assignments/blob/master/logistic_regression.ipynb)"
      ]
    },
    {
      "metadata": {
        "id": "JndnmDMp66FL",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Copyright 2017 Google LLC."
      ]
    },
    {
      "metadata": {
        "id": "hMqWDc_m6rUC",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#\n",
        "# https://www.apache.org/licenses/LICENSE-2.0\n",
        "#\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "g4T-_IsVbweU",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Logistic Regression"
      ]
    },
    {
      "metadata": {
        "id": "LEAHZv4rIYHX",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "**Learning Objectives:**\n",
        "  * Reframe the median house value predictor (from the preceding exercises) as a binary classification model\n",
        "  * Compare the effectiveness of logisitic regression vs linear regression for a binary classification problem"
      ]
    },
    {
      "metadata": {
        "id": "CnkCZqdIIYHY",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "As in the prior exercises, we're working with the [California housing data set](https://developers.google.com/machine-learning/crash-course/california-housing-data-description), but this time we will turn it into a binary classification problem by predicting whether a city block is a high-cost city block. We'll also revert to the default features, for now."
      ]
    },
    {
      "metadata": {
        "id": "9pltCyy2K3dd",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Frame the Problem as Binary Classification\n",
        "\n",
        "The target of our dataset is `median_house_value` which is a numeric (continuous-valued) feature. We can create a boolean label by applying a threshold to this continuous value.\n",
        "\n",
        "Given features describing a city block, we wish to predict if it is a high-cost city block. To prepare the targets for train and eval data, we define a classification threshold of the 75%-ile for median house value (a value of approximately 265000). All house values above the threshold are labeled `1`, and all others are labeled `0`."
      ]
    },
    {
      "metadata": {
        "id": "67IJwZX1Vvjt",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Setup\n",
        "\n",
        "Run the cells below to load the data and prepare the input features and targets."
      ]
    },
    {
      "metadata": {
        "id": "fOlbcJ4EIYHd",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from __future__ import print_function\n",
        "\n",
        "import math\n",
        "\n",
        "from IPython import display\n",
        "from matplotlib import cm\n",
        "from matplotlib import gridspec\n",
        "from matplotlib import pyplot as plt\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from sklearn import metrics\n",
        "import tensorflow as tf\n",
        "from tensorflow.python.data import Dataset\n",
        "\n",
        "tf.logging.set_verbosity(tf.logging.ERROR)\n",
        "pd.options.display.max_rows = 10\n",
        "pd.options.display.float_format = '{:.1f}'.format\n",
        "\n",
        "california_housing_dataframe = pd.read_csv(\"https://storage.googleapis.com/mledu-datasets/california_housing_train.csv\", sep=\",\")\n",
        "\n",
        "california_housing_dataframe = california_housing_dataframe.reindex(\n",
        "    np.random.permutation(california_housing_dataframe.index))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "lTB73MNeIYHf",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Note how the code below is slightly different from the previous exercises. Instead of using `median_house_value` as target, we create a new binary target, `median_house_value_is_high`."
      ]
    },
    {
      "metadata": {
        "id": "kPSqspaqIYHg",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def preprocess_features(california_housing_dataframe):\n",
        "  \"\"\"Prepares input features from California housing data set.\n",
        "\n",
        "  Args:\n",
        "    california_housing_dataframe: A Pandas DataFrame expected to contain data\n",
        "      from the California housing data set.\n",
        "  Returns:\n",
        "    A DataFrame that contains the features to be used for the model, including\n",
        "    synthetic features.\n",
        "  \"\"\"\n",
        "  selected_features = california_housing_dataframe[\n",
        "    [\"latitude\",\n",
        "     \"longitude\",\n",
        "     \"housing_median_age\",\n",
        "     \"total_rooms\",\n",
        "     \"total_bedrooms\",\n",
        "     \"population\",\n",
        "     \"households\",\n",
        "     \"median_income\"]]\n",
        "  processed_features = selected_features.copy()\n",
        "  # Create a synthetic feature.\n",
        "  processed_features[\"rooms_per_person\"] = (\n",
        "    california_housing_dataframe[\"total_rooms\"] /\n",
        "    california_housing_dataframe[\"population\"])\n",
        "  return processed_features\n",
        "\n",
        "def preprocess_targets(california_housing_dataframe):\n",
        "  \"\"\"Prepares target features (i.e., labels) from California housing data set.\n",
        "\n",
        "  Args:\n",
        "    california_housing_dataframe: A Pandas DataFrame expected to contain data\n",
        "      from the California housing data set.\n",
        "  Returns:\n",
        "    A DataFrame that contains the target feature.\n",
        "  \"\"\"\n",
        "  output_targets = pd.DataFrame()\n",
        "  # Create a boolean categorical feature representing whether the\n",
        "  # median_house_value is above a set threshold.\n",
        "  output_targets[\"median_house_value_is_high\"] = (\n",
        "    california_housing_dataframe[\"median_house_value\"] > 265000).astype(float)\n",
        "  return output_targets"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "FwOYWmXqWA6D",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1244
        },
        "outputId": "0a398f2a-a5a4-49d3-c8ec-992214b03a65"
      },
      "cell_type": "code",
      "source": [
        "# Choose the first 12000 (out of 17000) examples for training.\n",
        "training_examples = preprocess_features(california_housing_dataframe.head(12000))\n",
        "training_targets = preprocess_targets(california_housing_dataframe.head(12000))\n",
        "\n",
        "# Choose the last 5000 (out of 17000) examples for validation.\n",
        "validation_examples = preprocess_features(california_housing_dataframe.tail(5000))\n",
        "validation_targets = preprocess_targets(california_housing_dataframe.tail(5000))\n",
        "\n",
        "# Double-check that we've done the right thing.\n",
        "print(\"Training examples summary:\")\n",
        "display.display(training_examples.describe())\n",
        "print(\"Validation examples summary:\")\n",
        "display.display(validation_examples.describe())\n",
        "\n",
        "print(\"Training targets summary:\")\n",
        "display.display(training_targets.describe())\n",
        "print(\"Validation targets summary:\")\n",
        "display.display(validation_targets.describe())"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training examples summary:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "       latitude  longitude  housing_median_age  total_rooms  total_bedrooms  \\\n",
              "count   12000.0    12000.0             12000.0      12000.0         12000.0   \n",
              "mean       35.6     -119.6                28.6       2625.8           537.6   \n",
              "std         2.1        2.0                12.6       2144.0           418.6   \n",
              "min        32.5     -124.3                 1.0          2.0             1.0   \n",
              "25%        33.9     -121.8                18.0       1461.0           296.0   \n",
              "50%        34.2     -118.5                29.0       2120.5           432.0   \n",
              "75%        37.7     -118.0                37.0       3120.0           647.0   \n",
              "max        42.0     -114.3                52.0      32627.0          6445.0   \n",
              "\n",
              "       population  households  median_income  rooms_per_person  \n",
              "count     12000.0     12000.0        12000.0           12000.0  \n",
              "mean       1420.4       499.0            3.9               2.0  \n",
              "std        1108.5       380.5            1.9               1.2  \n",
              "min           3.0         1.0            0.5               0.1  \n",
              "25%         788.0       282.0            2.6               1.5  \n",
              "50%        1167.0       407.0            3.5               1.9  \n",
              "75%        1703.0       601.0            4.7               2.3  \n",
              "max       28566.0      6082.0           15.0              55.2  "
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>latitude</th>\n",
              "      <th>longitude</th>\n",
              "      <th>housing_median_age</th>\n",
              "      <th>total_rooms</th>\n",
              "      <th>total_bedrooms</th>\n",
              "      <th>population</th>\n",
              "      <th>households</th>\n",
              "      <th>median_income</th>\n",
              "      <th>rooms_per_person</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>35.6</td>\n",
              "      <td>-119.6</td>\n",
              "      <td>28.6</td>\n",
              "      <td>2625.8</td>\n",
              "      <td>537.6</td>\n",
              "      <td>1420.4</td>\n",
              "      <td>499.0</td>\n",
              "      <td>3.9</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>2.1</td>\n",
              "      <td>2.0</td>\n",
              "      <td>12.6</td>\n",
              "      <td>2144.0</td>\n",
              "      <td>418.6</td>\n",
              "      <td>1108.5</td>\n",
              "      <td>380.5</td>\n",
              "      <td>1.9</td>\n",
              "      <td>1.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>32.5</td>\n",
              "      <td>-124.3</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>33.9</td>\n",
              "      <td>-121.8</td>\n",
              "      <td>18.0</td>\n",
              "      <td>1461.0</td>\n",
              "      <td>296.0</td>\n",
              "      <td>788.0</td>\n",
              "      <td>282.0</td>\n",
              "      <td>2.6</td>\n",
              "      <td>1.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>34.2</td>\n",
              "      <td>-118.5</td>\n",
              "      <td>29.0</td>\n",
              "      <td>2120.5</td>\n",
              "      <td>432.0</td>\n",
              "      <td>1167.0</td>\n",
              "      <td>407.0</td>\n",
              "      <td>3.5</td>\n",
              "      <td>1.9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>37.7</td>\n",
              "      <td>-118.0</td>\n",
              "      <td>37.0</td>\n",
              "      <td>3120.0</td>\n",
              "      <td>647.0</td>\n",
              "      <td>1703.0</td>\n",
              "      <td>601.0</td>\n",
              "      <td>4.7</td>\n",
              "      <td>2.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>42.0</td>\n",
              "      <td>-114.3</td>\n",
              "      <td>52.0</td>\n",
              "      <td>32627.0</td>\n",
              "      <td>6445.0</td>\n",
              "      <td>28566.0</td>\n",
              "      <td>6082.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>55.2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Validation examples summary:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "       latitude  longitude  housing_median_age  total_rooms  total_bedrooms  \\\n",
              "count    5000.0     5000.0              5000.0       5000.0          5000.0   \n",
              "mean       35.6     -119.5                28.6       2686.4           543.8   \n",
              "std         2.1        2.0                12.6       2263.5           428.4   \n",
              "min        32.5     -124.3                 1.0         15.0             3.0   \n",
              "25%        33.9     -121.7                18.0       1464.8           297.0   \n",
              "50%        34.2     -118.5                28.0       2148.0           437.0   \n",
              "75%        37.7     -118.0                37.0       3209.0           653.0   \n",
              "max        42.0     -114.6                52.0      37937.0          5471.0   \n",
              "\n",
              "       population  households  median_income  rooms_per_person  \n",
              "count      5000.0      5000.0         5000.0            5000.0  \n",
              "mean       1451.6       506.5            3.9               2.0  \n",
              "std        1236.9       394.0            1.9               1.1  \n",
              "min          11.0         3.0            0.5               0.0  \n",
              "25%         792.8       281.0            2.6               1.5  \n",
              "50%        1168.0       412.5            3.6               1.9  \n",
              "75%        1768.0       612.2            4.8               2.3  \n",
              "max       35682.0      5189.0           15.0              52.0  "
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>latitude</th>\n",
              "      <th>longitude</th>\n",
              "      <th>housing_median_age</th>\n",
              "      <th>total_rooms</th>\n",
              "      <th>total_bedrooms</th>\n",
              "      <th>population</th>\n",
              "      <th>households</th>\n",
              "      <th>median_income</th>\n",
              "      <th>rooms_per_person</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>35.6</td>\n",
              "      <td>-119.5</td>\n",
              "      <td>28.6</td>\n",
              "      <td>2686.4</td>\n",
              "      <td>543.8</td>\n",
              "      <td>1451.6</td>\n",
              "      <td>506.5</td>\n",
              "      <td>3.9</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>2.1</td>\n",
              "      <td>2.0</td>\n",
              "      <td>12.6</td>\n",
              "      <td>2263.5</td>\n",
              "      <td>428.4</td>\n",
              "      <td>1236.9</td>\n",
              "      <td>394.0</td>\n",
              "      <td>1.9</td>\n",
              "      <td>1.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>32.5</td>\n",
              "      <td>-124.3</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>11.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>33.9</td>\n",
              "      <td>-121.7</td>\n",
              "      <td>18.0</td>\n",
              "      <td>1464.8</td>\n",
              "      <td>297.0</td>\n",
              "      <td>792.8</td>\n",
              "      <td>281.0</td>\n",
              "      <td>2.6</td>\n",
              "      <td>1.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>34.2</td>\n",
              "      <td>-118.5</td>\n",
              "      <td>28.0</td>\n",
              "      <td>2148.0</td>\n",
              "      <td>437.0</td>\n",
              "      <td>1168.0</td>\n",
              "      <td>412.5</td>\n",
              "      <td>3.6</td>\n",
              "      <td>1.9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>37.7</td>\n",
              "      <td>-118.0</td>\n",
              "      <td>37.0</td>\n",
              "      <td>3209.0</td>\n",
              "      <td>653.0</td>\n",
              "      <td>1768.0</td>\n",
              "      <td>612.2</td>\n",
              "      <td>4.8</td>\n",
              "      <td>2.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>42.0</td>\n",
              "      <td>-114.6</td>\n",
              "      <td>52.0</td>\n",
              "      <td>37937.0</td>\n",
              "      <td>5471.0</td>\n",
              "      <td>35682.0</td>\n",
              "      <td>5189.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>52.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Training targets summary:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "       median_house_value_is_high\n",
              "count                     12000.0\n",
              "mean                          0.3\n",
              "std                           0.4\n",
              "min                           0.0\n",
              "25%                           0.0\n",
              "50%                           0.0\n",
              "75%                           1.0\n",
              "max                           1.0"
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>median_house_value_is_high</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>12000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>0.4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Validation targets summary:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "       median_house_value_is_high\n",
              "count                      5000.0\n",
              "mean                          0.2\n",
              "std                           0.4\n",
              "min                           0.0\n",
              "25%                           0.0\n",
              "50%                           0.0\n",
              "75%                           0.0\n",
              "max                           1.0"
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>median_house_value_is_high</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>5000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>0.4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "uon1LB3A31VN",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## How Would Linear Regression Fare?\n",
        "To see why logistic regression is effective, let us first train a naive model that uses linear regression. This model will use labels with values in the set `{0, 1}` and will try to predict a continuous value that is as close as possible to `0` or `1`. Furthermore, we wish to interpret the output as a probability, so it would be ideal if the output will be within the range `(0, 1)`. We would then apply a threshold of `0.5` to determine the label.\n",
        "\n",
        "Run the cells below to train the linear regression model using [LinearRegressor](https://www.tensorflow.org/api_docs/python/tf/estimator/LinearRegressor)."
      ]
    },
    {
      "metadata": {
        "id": "smmUYRDtWOV_",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def construct_feature_columns(input_features):\n",
        "  \"\"\"Construct the TensorFlow Feature Columns.\n",
        "\n",
        "  Args:\n",
        "    input_features: The names of the numerical input features to use.\n",
        "  Returns:\n",
        "    A set of feature columns\n",
        "  \"\"\"\n",
        "  return set([tf.feature_column.numeric_column(my_feature)\n",
        "              for my_feature in input_features])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "B5OwSrr1yIKD",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def my_input_fn(features, targets, batch_size=1, shuffle=True, num_epochs=None):\n",
        "    \"\"\"Trains a linear regression model.\n",
        "  \n",
        "    Args:\n",
        "      features: pandas DataFrame of features\n",
        "      targets: pandas DataFrame of targets\n",
        "      batch_size: Size of batches to be passed to the model\n",
        "      shuffle: True or False. Whether to shuffle the data.\n",
        "      num_epochs: Number of epochs for which data should be repeated. None = repeat indefinitely\n",
        "    Returns:\n",
        "      Tuple of (features, labels) for next data batch\n",
        "    \"\"\"\n",
        "    \n",
        "    # Convert pandas data into a dict of np arrays.\n",
        "    features = {key:np.array(value) for key,value in dict(features).items()}                                            \n",
        " \n",
        "    # Construct a dataset, and configure batching/repeating.\n",
        "    ds = Dataset.from_tensor_slices((features,targets)) # warning: 2GB limit\n",
        "    ds = ds.batch(batch_size).repeat(num_epochs)\n",
        "    \n",
        "    # Shuffle the data, if specified.\n",
        "    if shuffle:\n",
        "      ds = ds.shuffle(10000)\n",
        "    \n",
        "    # Return the next batch of data.\n",
        "    features, labels = ds.make_one_shot_iterator().get_next()\n",
        "    return features, labels"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "SE2-hq8PIYHz",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def train_linear_regressor_model(\n",
        "    learning_rate,\n",
        "    steps,\n",
        "    batch_size,\n",
        "    training_examples,\n",
        "    training_targets,\n",
        "    validation_examples,\n",
        "    validation_targets):\n",
        "  \"\"\"Trains a linear regression model.\n",
        "  \n",
        "  In addition to training, this function also prints training progress information,\n",
        "  as well as a plot of the training and validation loss over time.\n",
        "  \n",
        "  Args:\n",
        "    learning_rate: A `float`, the learning rate.\n",
        "    steps: A non-zero `int`, the total number of training steps. A training step\n",
        "      consists of a forward and backward pass using a single batch.\n",
        "    batch_size: A non-zero `int`, the batch size.\n",
        "    training_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for training.\n",
        "    training_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for training.\n",
        "    validation_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for validation.\n",
        "    validation_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for validation.\n",
        "      \n",
        "  Returns:\n",
        "    A `LinearRegressor` object trained on the training data.\n",
        "  \"\"\"\n",
        "\n",
        "  periods = 10\n",
        "  steps_per_period = steps / periods\n",
        "\n",
        "  # Create a linear regressor object.\n",
        "  my_optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)\n",
        "  my_optimizer = tf.contrib.estimator.clip_gradients_by_norm(my_optimizer, 5.0)\n",
        "  linear_regressor = tf.estimator.LinearRegressor(\n",
        "      feature_columns=construct_feature_columns(training_examples),\n",
        "      optimizer=my_optimizer\n",
        "  )\n",
        "    \n",
        "  # Create input functions.\n",
        "  training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                          training_targets[\"median_house_value_is_high\"], \n",
        "                                          batch_size=batch_size)\n",
        "  predict_training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                                  training_targets[\"median_house_value_is_high\"], \n",
        "                                                  num_epochs=1, \n",
        "                                                  shuffle=False)\n",
        "  predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                    validation_targets[\"median_house_value_is_high\"], \n",
        "                                                    num_epochs=1, \n",
        "                                                    shuffle=False)\n",
        "\n",
        "  # Train the model, but do so inside a loop so that we can periodically assess\n",
        "  # loss metrics.\n",
        "  print(\"Training model...\")\n",
        "  print(\"RMSE (on training data):\")\n",
        "  training_rmse = []\n",
        "  validation_rmse = []\n",
        "  for period in range (0, periods):\n",
        "    # Train the model, starting from the prior state.\n",
        "    linear_regressor.train(\n",
        "        input_fn=training_input_fn,\n",
        "        steps=steps_per_period\n",
        "    )\n",
        "    \n",
        "    # Take a break and compute predictions.\n",
        "    training_predictions = linear_regressor.predict(input_fn=predict_training_input_fn)\n",
        "    training_predictions = np.array([item['predictions'][0] for item in training_predictions])\n",
        "    \n",
        "    validation_predictions = linear_regressor.predict(input_fn=predict_validation_input_fn)\n",
        "    validation_predictions = np.array([item['predictions'][0] for item in validation_predictions])\n",
        "    \n",
        "    # Compute training and validation loss.\n",
        "    training_root_mean_squared_error = math.sqrt(\n",
        "        metrics.mean_squared_error(training_predictions, training_targets))\n",
        "    validation_root_mean_squared_error = math.sqrt(\n",
        "        metrics.mean_squared_error(validation_predictions, validation_targets))\n",
        "    # Occasionally print the current loss.\n",
        "    print(\"  period %02d : %0.2f\" % (period, training_root_mean_squared_error))\n",
        "    # Add the loss metrics from this period to our list.\n",
        "    training_rmse.append(training_root_mean_squared_error)\n",
        "    validation_rmse.append(validation_root_mean_squared_error)\n",
        "  print(\"Model training finished.\")\n",
        "  \n",
        "  # Output a graph of loss metrics over periods.\n",
        "  plt.ylabel(\"RMSE\")\n",
        "  plt.xlabel(\"Periods\")\n",
        "  plt.title(\"Root Mean Squared Error vs. Periods\")\n",
        "  plt.tight_layout()\n",
        "  plt.plot(training_rmse, label=\"training\")\n",
        "  plt.plot(validation_rmse, label=\"validation\")\n",
        "  plt.legend()\n",
        "\n",
        "  return linear_regressor"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "TDBD8xeeIYH2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 646
        },
        "outputId": "6c1b4bcc-d219-40eb-dee7-7be317d40d97"
      },
      "cell_type": "code",
      "source": [
        "linear_regressor = train_linear_regressor_model(\n",
        "    learning_rate=0.0001,\n",
        "    steps=100,\n",
        "    batch_size=10,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training model...\n",
            "RMSE (on training data):\n",
            "  period 00 : 1.77\n",
            "  period 01 : 1.72\n",
            "  period 02 : 1.63\n",
            "  period 03 : 1.63\n",
            "  period 04 : 1.62\n",
            "  period 05 : 1.62\n",
            "  period 06 : 1.61\n",
            "  period 07 : 1.60\n",
            "  period 08 : 1.67\n",
            "  period 09 : 1.69\n",
            "Model training finished.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjMAAAGACAYAAABY5OOEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f13516e8990>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "JjBZ_q7aD9gh",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Task 1: Can We Calculate LogLoss for These Predictions?\n",
        "\n",
        "**Examine the predictions and decide whether or not we can use them to calculate LogLoss.**\n",
        "\n",
        "`LinearRegressor` uses the L2 loss, which doesn't do a great job at penalizing misclassifications when the output is interpreted as a probability.  For example, there should be a huge difference whether a negative example is classified as positive with a probability of 0.9 vs 0.9999, but L2 loss doesn't strongly differentiate these cases.\n",
        "\n",
        "In contrast, `LogLoss` penalizes these \"confidence errors\" much more heavily.  Remember, `LogLoss` is defined as:\n",
        "\n",
        "$$Log Loss = \\sum_{(x,y)\\in D} -y \\cdot log(y_{pred}) - (1 - y) \\cdot log(1 - y_{pred})$$\n",
        "\n",
        "\n",
        "But first, we'll need to obtain the prediction values. We could use `LinearRegressor.predict` to obtain these.\n",
        "\n",
        "Given the predictions and that targets, can we calculate `LogLoss`?"
      ]
    },
    {
      "metadata": {
        "id": "rYwsdfgstoyr",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 460
        },
        "outputId": "f766a5f7-5216-4ad3-f8e3-360c01730a06"
      },
      "cell_type": "code",
      "source": [
        "predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                    validation_targets[\"median_house_value_is_high\"], \n",
        "                                                    num_epochs=1, \n",
        "                                                    shuffle=False)\n",
        "validation_predictions = linear_regressor.predict(input_fn=predict_validation_input_fn)\n",
        "validation_predictions = np.array([item['predictions'][0] for item in validation_predictions])\n",
        "\n",
        "plt.hist(validation_predictions)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([4.087e+03, 7.420e+02, 1.230e+02, 2.700e+01, 1.100e+01, 3.000e+00,\n",
              "        2.000e+00, 3.000e+00, 1.000e+00, 1.000e+00]),\n",
              " array([ 0.02738386,  2.03673673,  4.0460896 ,  6.05544247,  8.06479534,\n",
              "        10.07414821, 12.08350108, 14.09285395, 16.10220681, 18.11155968,\n",
              "        20.12091255]),\n",
              " <a list of 10 Patch objects>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAFKCAYAAADScRzUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYBQ1AAAGo6gBADAYRQ0AgMH+P00dRsffgp2FAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f1377e5a610>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "dPpJUV862FYI",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Solution\n",
        "\n",
        "Click below to display the solution."
      ]
    },
    {
      "metadata": {
        "id": "kXFQ5uig2RoP",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                  validation_targets[\"median_house_value_is_high\"], \n",
        "                                                  num_epochs=1, \n",
        "                                                  shuffle=False)\n",
        "\n",
        "validation_predictions = linear_regressor.predict(input_fn=predict_validation_input_fn)\n",
        "validation_predictions = np.array([item['predictions'][0] for item in validation_predictions])\n",
        "\n",
        "_ = plt.hist(validation_predictions)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "rYpy336F9wBg",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Task 2: Train a Logistic Regression Model and Calculate LogLoss on the Validation Set\n",
        "\n",
        "To use logistic regression, simply use [LinearClassifier](https://www.tensorflow.org/api_docs/python/tf/estimator/LinearClassifier) instead of `LinearRegressor`. Complete the code below.\n",
        "\n",
        "**NOTE**: When running `train()` and `predict()` on a `LinearClassifier` model, you can access the real-valued predicted probabilities via the `\"probabilities\"` key in the returned dict—e.g., `predictions[\"probabilities\"]`. Sklearn's [log_loss](http://scikit-learn.org/stable/modules/generated/sklearn.metrics.log_loss.html) function is handy for calculating LogLoss using these probabilities.\n"
      ]
    },
    {
      "metadata": {
        "id": "JElcb--E9wBm",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def train_linear_classifier_model(\n",
        "    learning_rate,\n",
        "    steps,\n",
        "    batch_size,\n",
        "    training_examples,\n",
        "    training_targets,\n",
        "    validation_examples,\n",
        "    validation_targets):\n",
        "  \"\"\"Trains a linear classification model.\n",
        "  \n",
        "  In addition to training, this function also prints training progress information,\n",
        "  as well as a plot of the training and validation loss over time.\n",
        "  \n",
        "  Args:\n",
        "    learning_rate: A `float`, the learning rate.\n",
        "    steps: A non-zero `int`, the total number of training steps. A training step\n",
        "      consists of a forward and backward pass using a single batch.\n",
        "    batch_size: A non-zero `int`, the batch size.\n",
        "    training_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for training.\n",
        "    training_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for training.\n",
        "    validation_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for validation.\n",
        "    validation_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for validation.\n",
        "      \n",
        "  Returns:\n",
        "    A `LinearClassifier` object trained on the training data.\n",
        "  \"\"\"\n",
        "\n",
        "  periods = 10\n",
        "  steps_per_period = steps / periods\n",
        "  \n",
        "  # Create a linear classifier object.\n",
        "  my_optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)\n",
        "  my_optimizer = tf.contrib.estimator.clip_gradients_by_norm(my_optimizer, 5.0)\n",
        "  linear_classifier = tf.estimator.LinearClassifier(\n",
        "      feature_columns=construct_feature_columns(training_examples),\n",
        "      optimizer=my_optimizer\n",
        "  )\n",
        "  \n",
        "  # Create input functions.\n",
        "  training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                          training_targets[\"median_house_value_is_high\"], \n",
        "                                          batch_size=batch_size)\n",
        "  predict_training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                                  training_targets[\"median_house_value_is_high\"], \n",
        "                                                  num_epochs=1, \n",
        "                                                  shuffle=False)\n",
        "  predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                    validation_targets[\"median_house_value_is_high\"], \n",
        "                                                    num_epochs=1, \n",
        "                                                    shuffle=False)\n",
        "  \n",
        "  # Train the model, but do so inside a loop so that we can periodically assess\n",
        "  # loss metrics.\n",
        "  print(\"Training model...\")\n",
        "  print(\"LogLoss (on training data):\")\n",
        "  training_log_losses = []\n",
        "  validation_log_losses = []\n",
        "  for period in range (0, periods):\n",
        "    # Train the model, starting from the prior state.\n",
        "    linear_classifier.train(\n",
        "        input_fn=training_input_fn,\n",
        "        steps=steps_per_period\n",
        "    )\n",
        "    # Take a break and compute predictions.    \n",
        "    training_probabilities = linear_classifier.predict(input_fn=predict_training_input_fn)\n",
        "    training_probabilities = np.array([item['probabilities'] for item in training_probabilities])\n",
        "    \n",
        "    validation_probabilities = linear_classifier.predict(input_fn=predict_validation_input_fn)\n",
        "    validation_probabilities = np.array([item['probabilities'] for item in validation_probabilities])\n",
        "    \n",
        "    training_log_loss = metrics.log_loss(training_targets, training_probabilities)\n",
        "    validation_log_loss = metrics.log_loss(validation_targets, validation_probabilities)\n",
        "    # Occasionally print the current loss.\n",
        "    print(\"  period %02d : %0.2f\" % (period, training_log_loss))\n",
        "    # Add the loss metrics from this period to our list.\n",
        "    training_log_losses.append(training_log_loss)\n",
        "    validation_log_losses.append(validation_log_loss)\n",
        "  print(\"Model training finished.\")\n",
        "  \n",
        "  # Output a graph of loss metrics over periods.\n",
        "  plt.ylabel(\"LogLoss\")\n",
        "  plt.xlabel(\"Periods\")\n",
        "  plt.title(\"LogLoss vs. Periods\")\n",
        "  plt.tight_layout()\n",
        "  plt.plot(training_log_losses, label=\"training\")\n",
        "  plt.plot(validation_log_losses, label=\"validation\")\n",
        "  plt.legend()\n",
        "\n",
        "  return linear_classifier"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "VM0wmnFUIYH9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 646
        },
        "outputId": "4849e87f-de3e-4659-9a2e-a88571618012"
      },
      "cell_type": "code",
      "source": [
        "linear_classifier = train_linear_classifier_model(\n",
        "    learning_rate=0.000005,\n",
        "    steps=500,\n",
        "    batch_size=20,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training model...\n",
            "LogLoss (on training data):\n",
            "  period 00 : 0.61\n",
            "  period 01 : 0.58\n",
            "  period 02 : 0.57\n",
            "  period 03 : 0.56\n",
            "  period 04 : 0.55\n",
            "  period 05 : 0.54\n",
            "  period 06 : 0.54\n",
            "  period 07 : 0.54\n",
            "  period 08 : 0.53\n",
            "  period 09 : 0.53\n",
            "Model training finished.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjMAAAGACAYAAABY5OOEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f1377e18d10>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "i2e3TlyL57Qs",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Solution\n",
        "\n",
        "Click below to see the solution.\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "5YxXd2hn6MuF",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def train_linear_classifier_model(\n",
        "    learning_rate,\n",
        "    steps,\n",
        "    batch_size,\n",
        "    training_examples,\n",
        "    training_targets,\n",
        "    validation_examples,\n",
        "    validation_targets):\n",
        "  \"\"\"Trains a linear classification model.\n",
        "  \n",
        "  In addition to training, this function also prints training progress information,\n",
        "  as well as a plot of the training and validation loss over time.\n",
        "  \n",
        "  Args:\n",
        "    learning_rate: A `float`, the learning rate.\n",
        "    steps: A non-zero `int`, the total number of training steps. A training step\n",
        "      consists of a forward and backward pass using a single batch.\n",
        "    batch_size: A non-zero `int`, the batch size.\n",
        "    training_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for training.\n",
        "    training_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for training.\n",
        "    validation_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for validation.\n",
        "    validation_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for validation.\n",
        "      \n",
        "  Returns:\n",
        "    A `LinearClassifier` object trained on the training data.\n",
        "  \"\"\"\n",
        "\n",
        "  periods = 10\n",
        "  steps_per_period = steps / periods\n",
        "  \n",
        "  # Create a linear classifier object.\n",
        "  my_optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)\n",
        "  my_optimizer = tf.contrib.estimator.clip_gradients_by_norm(my_optimizer, 5.0)  \n",
        "  linear_classifier = tf.estimator.LinearClassifier(\n",
        "      feature_columns=construct_feature_columns(training_examples),\n",
        "      optimizer=my_optimizer\n",
        "  )\n",
        "  \n",
        "  # Create input functions.\n",
        "  training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                          training_targets[\"median_house_value_is_high\"], \n",
        "                                          batch_size=batch_size)\n",
        "  predict_training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                                  training_targets[\"median_house_value_is_high\"], \n",
        "                                                  num_epochs=1, \n",
        "                                                  shuffle=False)\n",
        "  predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                    validation_targets[\"median_house_value_is_high\"], \n",
        "                                                    num_epochs=1, \n",
        "                                                    shuffle=False)\n",
        "  \n",
        "  # Train the model, but do so inside a loop so that we can periodically assess\n",
        "  # loss metrics.\n",
        "  print(\"Training model...\")\n",
        "  print(\"LogLoss (on training data):\")\n",
        "  training_log_losses = []\n",
        "  validation_log_losses = []\n",
        "  for period in range (0, periods):\n",
        "    # Train the model, starting from the prior state.\n",
        "    linear_classifier.train(\n",
        "        input_fn=training_input_fn,\n",
        "        steps=steps_per_period\n",
        "    )\n",
        "    # Take a break and compute predictions.    \n",
        "    training_probabilities = linear_classifier.predict(input_fn=predict_training_input_fn)\n",
        "    training_probabilities = np.array([item['probabilities'] for item in training_probabilities])\n",
        "    \n",
        "    validation_probabilities = linear_classifier.predict(input_fn=predict_validation_input_fn)\n",
        "    validation_probabilities = np.array([item['probabilities'] for item in validation_probabilities])\n",
        "    \n",
        "    training_log_loss = metrics.log_loss(training_targets, training_probabilities)\n",
        "    validation_log_loss = metrics.log_loss(validation_targets, validation_probabilities)\n",
        "    # Occasionally print the current loss.\n",
        "    print(\"  period %02d : %0.2f\" % (period, training_log_loss))\n",
        "    # Add the loss metrics from this period to our list.\n",
        "    training_log_losses.append(training_log_loss)\n",
        "    validation_log_losses.append(validation_log_loss)\n",
        "  print(\"Model training finished.\")\n",
        "  \n",
        "  # Output a graph of loss metrics over periods.\n",
        "  plt.ylabel(\"LogLoss\")\n",
        "  plt.xlabel(\"Periods\")\n",
        "  plt.title(\"LogLoss vs. Periods\")\n",
        "  plt.tight_layout()\n",
        "  plt.plot(training_log_losses, label=\"training\")\n",
        "  plt.plot(validation_log_losses, label=\"validation\")\n",
        "  plt.legend()\n",
        "\n",
        "  return linear_classifier"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "UPM_T1FXsTaL",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "linear_classifier = train_linear_classifier_model(\n",
        "    learning_rate=0.000005,\n",
        "    steps=500,\n",
        "    batch_size=20,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "i-Xo83_aR6s_",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Task 3: Calculate Accuracy and plot a ROC Curve for the Validation Set\n",
        "\n",
        "A few of the metrics useful for classification are the model [accuracy](https://en.wikipedia.org/wiki/Accuracy_and_precision#In_binary_classification), the [ROC curve](https://en.wikipedia.org/wiki/Receiver_operating_characteristic) and the area under the ROC curve (AUC). We'll examine these metrics.\n",
        "\n",
        "`LinearClassifier.evaluate` calculates useful metrics like accuracy and AUC."
      ]
    },
    {
      "metadata": {
        "id": "DKSQ87VVIYIA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 55
        },
        "outputId": "ceae5b51-7a6f-4e9a-fd85-a114accf3256"
      },
      "cell_type": "code",
      "source": [
        "evaluation_metrics = linear_classifier.evaluate(input_fn=predict_validation_input_fn)\n",
        "\n",
        "print(\"AUC on the validation set: %0.2f\" % evaluation_metrics['auc'])\n",
        "print(\"Accuracy on the validation set: %0.2f\" % evaluation_metrics['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "AUC on the validation set: 0.72\n",
            "Accuracy on the validation set: 0.75\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "47xGS2uNIYIE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "You may use class probabilities, such as those calculated by `LinearClassifier.predict`,\n",
        "and Sklearn's [roc_curve](http://scikit-learn.org/stable/modules/model_evaluation.html#roc-metrics) to\n",
        "obtain the true positive and false positive rates needed to plot a ROC curve."
      ]
    },
    {
      "metadata": {
        "id": "xaU7ttj8IYIF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 347
        },
        "outputId": "207cb69e-34d4-4e26-913a-a4dde24307d8"
      },
      "cell_type": "code",
      "source": [
        "validation_probabilities = linear_classifier.predict(input_fn=predict_validation_input_fn)\n",
        "# Get just the probabilities for the positive class.\n",
        "validation_probabilities = np.array([item['probabilities'][1] for item in validation_probabilities])\n",
        "\n",
        "false_positive_rate, true_positive_rate, thresholds = metrics.roc_curve(\n",
        "    validation_targets, validation_probabilities)\n",
        "plt.plot(false_positive_rate, true_positive_rate, label=\"our model\")\n",
        "plt.plot([0, 1], [0, 1], label=\"random classifier\")\n",
        "_ = plt.legend(loc=2)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeEAAAFKCAYAAAAqkecjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwiIiIiZRCYuIiJhEJSwiImKS/wNSqicb578m2wAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f1377bd8a50>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "PIdhwfgzIYII",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "**See if you can tune the learning settings of the model trained at Task 2 to improve AUC.**\n",
        "\n",
        "Often times, certain metrics improve at the detriment of others, and you'll need to find the settings that achieve a good compromise.\n",
        "\n",
        "**Verify if all metrics improve at the same time.**"
      ]
    },
    {
      "metadata": {
        "id": "XKIqjsqcCaxO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 684
        },
        "outputId": "d375492d-ec90-4ec4-c844-a253c9ba5d04"
      },
      "cell_type": "code",
      "source": [
        "# TUNE THE SETTINGS BELOW TO IMPROVE AUC\n",
        "linear_classifier = train_linear_classifier_model(\n",
        "    learning_rate=0.000005,\n",
        "    steps=10000,\n",
        "    batch_size=50,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)\n",
        "\n",
        "evaluation_metrics = linear_classifier.evaluate(input_fn=predict_validation_input_fn)\n",
        "\n",
        "print(\"AUC on the validation set: %0.2f\" % evaluation_metrics['auc'])\n",
        "print(\"Accuracy on the validation set: %0.2f\" % evaluation_metrics['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training model...\n",
            "LogLoss (on training data):\n",
            "  period 00 : 0.52\n",
            "  period 01 : 0.51\n",
            "  period 02 : 0.50\n",
            "  period 03 : 0.50\n",
            "  period 04 : 0.49\n",
            "  period 05 : 0.49\n",
            "  period 06 : 0.48\n",
            "  period 07 : 0.48\n",
            "  period 08 : 0.49\n",
            "  period 09 : 0.48\n",
            "Model training finished.\n",
            "AUC on the validation set: 0.79\n",
            "Accuracy on the validation set: 0.78\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjMAAAGACAYAAABY5OOEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWqDaAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f1350182f50>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "wCugvl0JdWYL",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Solution\n",
        "\n",
        "Click below for a possible solution."
      ]
    },
    {
      "metadata": {
        "id": "VHosS1g2aetf",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "One possible solution that works is to just train for longer, as long as we don't overfit. \n",
        "\n",
        "We can do this by increasing the number the steps, the batch size, or both.\n",
        "\n",
        "All metrics improve at the same time, so our loss metric is a good proxy\n",
        "for both AUC and accuracy.\n",
        "\n",
        "Notice how it takes many, many more iterations just to squeeze a few more \n",
        "units of AUC. This commonly happens. But often even this small gain is worth \n",
        "the costs."
      ]
    },
    {
      "metadata": {
        "id": "dWgTEYMddaA-",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "linear_classifier = train_linear_classifier_model(\n",
        "    learning_rate=0.000003,\n",
        "    steps=20000,\n",
        "    batch_size=500,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)\n",
        "\n",
        "evaluation_metrics = linear_classifier.evaluate(input_fn=predict_validation_input_fn)\n",
        "\n",
        "print(\"AUC on the validation set: %0.2f\" % evaluation_metrics['auc'])\n",
        "print(\"Accuracy on the validation set: %0.2f\" % evaluation_metrics['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}