classification-regression.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/bashtavenko/colabs/blob/main/classification-regression.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "q18CH71nyUof"
      },
      "source": [
        "This is a companion notebook for the book [Deep Learning with Python, Second Edition](https://www.manning.com/books/deep-learning-with-python-second-edition?a_aid=keras&a_bid=76564dff). For readability, it only contains runnable code blocks and section titles, and omits everything else in the book: text paragraphs, figures, and pseudocode.\n",
        "\n",
        "**If you want to be able to follow what's going on, I recommend reading the notebook side by side with your copy of the book.**\n",
        "\n",
        "This notebook was generated for TensorFlow 2.6."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MOIcRQhlyUog"
      },
      "source": [
        "# Getting started with neural networks: Classification and regression"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EONxxXHTyUoh"
      },
      "source": [
        "## Classifying movie reviews: A binary classification example"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rwQbObIwyUoh"
      },
      "source": [
        "### The IMDB dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x45ZqcCLyUoi"
      },
      "source": [
        "**Loading the IMDB dataset**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uCjEL7hCyUoi"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.datasets import imdb\n",
        "(train_data, train_labels), (test_data, test_labels) = imdb.load_data(\n",
        "    num_words=10000)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QVnR4xdAyUoj"
      },
      "outputs": [],
      "source": [
        "train_data[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4TMoaaUmyUoj"
      },
      "outputs": [],
      "source": [
        "train_labels[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4X5N8cFYyUoj"
      },
      "outputs": [],
      "source": [
        "max([max(sequence) for sequence in train_data])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "12dVqi3ayUok"
      },
      "source": [
        "**Decoding reviews back to text**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bc70pER3yUok"
      },
      "outputs": [],
      "source": [
        "word_index = imdb.get_word_index()\n",
        "reverse_word_index = dict(\n",
        "    [(value, key) for (key, value) in word_index.items()])\n",
        "decoded_review = \" \".join(\n",
        "    [reverse_word_index.get(i - 3, \"?\") for i in train_data[0]])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "On8ktMLNyUok"
      },
      "source": [
        "### Preparing the data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XhbY1oG8yUok"
      },
      "source": [
        "**Encoding the integer sequences via multi-hot encoding**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4s5OYy4zyUok"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "def vectorize_sequences(sequences, dimension=10000):\n",
        "    results = np.zeros((len(sequences), dimension))\n",
        "    for i, sequence in enumerate(sequences):\n",
        "        for j in sequence:\n",
        "            results[i, j] = 1.\n",
        "    return results\n",
        "x_train = vectorize_sequences(train_data)\n",
        "x_test = vectorize_sequences(test_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7weK4V2tyUok"
      },
      "outputs": [],
      "source": [
        "x_train[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "AyfG_cYlyUol"
      },
      "outputs": [],
      "source": [
        "y_train = np.asarray(train_labels).astype(\"float32\")\n",
        "y_test = np.asarray(test_labels).astype(\"float32\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nOZYS__lyUol"
      },
      "source": [
        "### Building your model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FZtr6TwCyUol"
      },
      "source": [
        "**Model definition**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2H9qSx-nyUol"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "49nyWGdSyUol"
      },
      "source": [
        "**Compiling the model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OgPrfxWvyUol"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AjTz6rf5yUom"
      },
      "source": [
        "### Validating your approach"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l17pJ811yUom"
      },
      "source": [
        "**Setting aside a validation set**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Q3kk-wSHyUom"
      },
      "outputs": [],
      "source": [
        "x_val = x_train[:10000]\n",
        "partial_x_train = x_train[10000:]\n",
        "y_val = y_train[:10000]\n",
        "partial_y_train = y_train[10000:]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H2wLUP70yUom"
      },
      "source": [
        "**Training your model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zWH5rRXcyUom"
      },
      "outputs": [],
      "source": [
        "history = model.fit(partial_x_train,\n",
        "                    partial_y_train,\n",
        "                    epochs=20,\n",
        "                    batch_size=512,\n",
        "                    validation_data=(x_val, y_val))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jXX6bv-yyUom"
      },
      "outputs": [],
      "source": [
        "history_dict = history.history\n",
        "history_dict.keys()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9ERUebgYyUom"
      },
      "source": [
        "**Plotting the training and validation loss**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0_gSPEpUyUon"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "history_dict = history.history\n",
        "loss_values = history_dict[\"loss\"]\n",
        "val_loss_values = history_dict[\"val_loss\"]\n",
        "epochs = range(1, len(loss_values) + 1)\n",
        "plt.plot(epochs, loss_values, \"bo\", label=\"Training loss\")\n",
        "plt.plot(epochs, val_loss_values, \"b\", label=\"Validation loss\")\n",
        "plt.title(\"Training and validation loss\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3f5n7rU0yUon"
      },
      "source": [
        "**Plotting the training and validation accuracy**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "O2L6rc2YyUon"
      },
      "outputs": [],
      "source": [
        "plt.clf()\n",
        "acc = history_dict[\"accuracy\"]\n",
        "val_acc = history_dict[\"val_accuracy\"]\n",
        "plt.plot(epochs, acc, \"bo\", label=\"Training acc\")\n",
        "plt.plot(epochs, val_acc, \"b\", label=\"Validation acc\")\n",
        "plt.title(\"Training and validation accuracy\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7pVFXgiCyUon"
      },
      "source": [
        "**Retraining a model from scratch**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vJmQmngKyUon"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(x_train, y_train, epochs=4, batch_size=512)\n",
        "results = model.evaluate(x_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "O1ietkoJyUoo"
      },
      "outputs": [],
      "source": [
        "results"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iTtfpQspyUoo"
      },
      "source": [
        "### Using a trained model to generate predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bQTfjP3EyUoo"
      },
      "outputs": [],
      "source": [
        "model.predict(x_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-zN189nzyUoo"
      },
      "source": [
        "### Further experiments"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OGUQfblDyUoo"
      },
      "source": [
        "### Wrapping up"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CQfrl2O_yUoo"
      },
      "source": [
        "## Classifying newswires: A multiclass classification example"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qtySz7gyyUoo"
      },
      "source": [
        "### The Reuters dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Y9m5NxrJyUoo"
      },
      "source": [
        "**Loading the Reuters dataset**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ibuq0zdoyUop"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.datasets import reuters\n",
        "(train_data, train_labels), (test_data, test_labels) = reuters.load_data(\n",
        "    num_words=10000)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "H0bkhmNgyUop"
      },
      "outputs": [],
      "source": [
        "len(train_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ma4w06ISyUop"
      },
      "outputs": [],
      "source": [
        "len(test_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "syCQp86myUop"
      },
      "outputs": [],
      "source": [
        "train_data[10]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "T5iFpU0syUop"
      },
      "source": [
        "**Decoding newswires back to text**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FiN3DRJhyUop"
      },
      "outputs": [],
      "source": [
        "word_index = reuters.get_word_index()\n",
        "reverse_word_index = dict([(value, key) for (key, value) in word_index.items()])\n",
        "decoded_newswire = \" \".join([reverse_word_index.get(i - 3, \"?\") for i in\n",
        "    train_data[0]])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wdVpk8MEyUop"
      },
      "outputs": [],
      "source": [
        "train_labels[10]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E-RrBeHkyUop"
      },
      "source": [
        "### Preparing the data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uJJQxxjVyUoq"
      },
      "source": [
        "**Encoding the input data**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8UtWJIgAyUoq"
      },
      "outputs": [],
      "source": [
        "x_train = vectorize_sequences(train_data)\n",
        "x_test = vectorize_sequences(test_data)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J7SG1ObpyUoq"
      },
      "source": [
        "**Encoding the labels**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7nGUIU-MyUoq"
      },
      "outputs": [],
      "source": [
        "def to_one_hot(labels, dimension=46):\n",
        "    results = np.zeros((len(labels), dimension))\n",
        "    for i, label in enumerate(labels):\n",
        "        results[i, label] = 1.\n",
        "    return results\n",
        "y_train = to_one_hot(train_labels)\n",
        "y_test = to_one_hot(test_labels)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "nFZKiAqbyUoq"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.utils import to_categorical\n",
        "y_train = to_categorical(train_labels)\n",
        "y_test = to_categorical(test_labels)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7n_rjPZUyUoq"
      },
      "source": [
        "### Building your model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RNVuCtfpyUoq"
      },
      "source": [
        "**Model definition**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "nlgNX557yUor"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(46, activation=\"softmax\")\n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FZZZlCG3yUor"
      },
      "source": [
        "**Compiling the model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0yxOmGr9yUor"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6OAN3L23yUor"
      },
      "source": [
        "### Validating your approach"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e_TOrFkzyUor"
      },
      "source": [
        "**Setting aside a validation set**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gvVTUMl3yUor"
      },
      "outputs": [],
      "source": [
        "x_val = x_train[:1000]\n",
        "partial_x_train = x_train[1000:]\n",
        "y_val = y_train[:1000]\n",
        "partial_y_train = y_train[1000:]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2f8Qmmx5yUor"
      },
      "source": [
        "**Training the model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SAaY9tVyyUor"
      },
      "outputs": [],
      "source": [
        "history = model.fit(partial_x_train,\n",
        "                    partial_y_train,\n",
        "                    epochs=20,\n",
        "                    batch_size=512,\n",
        "                    validation_data=(x_val, y_val))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c_Hwa-BoyUos"
      },
      "source": [
        "**Plotting the training and validation loss**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "AC7URwsFyUos"
      },
      "outputs": [],
      "source": [
        "loss = history.history[\"loss\"]\n",
        "val_loss = history.history[\"val_loss\"]\n",
        "epochs = range(1, len(loss) + 1)\n",
        "plt.plot(epochs, loss, \"bo\", label=\"Training loss\")\n",
        "plt.plot(epochs, val_loss, \"b\", label=\"Validation loss\")\n",
        "plt.title(\"Training and validation loss\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oAL7G5_8yUos"
      },
      "source": [
        "**Plotting the training and validation accuracy**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "YgedvVCLyUos"
      },
      "outputs": [],
      "source": [
        "plt.clf()\n",
        "acc = history.history[\"accuracy\"]\n",
        "val_acc = history.history[\"val_accuracy\"]\n",
        "plt.plot(epochs, acc, \"bo\", label=\"Training accuracy\")\n",
        "plt.plot(epochs, val_acc, \"b\", label=\"Validation accuracy\")\n",
        "plt.title(\"Training and validation accuracy\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3QvBQZxjyUos"
      },
      "source": [
        "**Retraining a model from scratch**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kGpqUYqTyUos"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([\n",
        "  layers.Dense(64, activation=\"relu\"),\n",
        "  layers.Dense(64, activation=\"relu\"),\n",
        "  layers.Dense(46, activation=\"softmax\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(x_train,\n",
        "          y_train,\n",
        "          epochs=9,\n",
        "          batch_size=512)\n",
        "results = model.evaluate(x_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "MOIsl2YZyUos"
      },
      "outputs": [],
      "source": [
        "results"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZKby1yF1yUot"
      },
      "outputs": [],
      "source": [
        "import copy\n",
        "test_labels_copy = copy.copy(test_labels)\n",
        "np.random.shuffle(test_labels_copy)\n",
        "hits_array = np.array(test_labels) == np.array(test_labels_copy)\n",
        "hits_array.mean()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7U0ZATVWyUot"
      },
      "source": [
        "### Generating predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WIEDW7DcyUot"
      },
      "outputs": [],
      "source": [
        "predictions = model.predict(x_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-vJdnbICyUot"
      },
      "outputs": [],
      "source": [
        "predictions[0].shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "6SUiXuNfyUot"
      },
      "outputs": [],
      "source": [
        "np.sum(predictions[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jGPzU2IQyUot"
      },
      "outputs": [],
      "source": [
        "np.argmax(predictions[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JOAqvwo0yUot"
      },
      "source": [
        "### A different way to handle the labels and the loss"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TQATBpHmyUot"
      },
      "outputs": [],
      "source": [
        "y_train = np.array(train_labels)\n",
        "y_test = np.array(test_labels)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XCqimnSSyUou"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o3AfcotgyUou"
      },
      "source": [
        "### The importance of having sufficiently large intermediate layers"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IIODwIdDyUou"
      },
      "source": [
        "**A model with an information bottleneck**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "psaefDe0yUou"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(4, activation=\"relu\"),\n",
        "    layers.Dense(46, activation=\"softmax\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(partial_x_train,\n",
        "          partial_y_train,\n",
        "          epochs=20,\n",
        "          batch_size=128,\n",
        "          validation_data=(x_val, y_val))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SXjoll6RyUou"
      },
      "source": [
        "### Further experiments"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BSoKn6nMyUou"
      },
      "source": [
        "### Wrapping up"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z-gPkFnXyUou"
      },
      "source": [
        "## Predicting house prices: A regression example"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jsqwb6boyUou"
      },
      "source": [
        "### The Boston Housing Price dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wiCPwFIGyUov"
      },
      "source": [
        "**Loading the Boston housing dataset**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-eszdWrUyUov"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.datasets import boston_housing\n",
        "(train_data, train_targets), (test_data, test_targets) = boston_housing.load_data()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IXnn3zY8yUov"
      },
      "outputs": [],
      "source": [
        "train_data.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VFzJiyDHyUov"
      },
      "outputs": [],
      "source": [
        "test_data.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0fo3P2BlyUov"
      },
      "outputs": [],
      "source": [
        "train_targets"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7VmhehY7yUov"
      },
      "source": [
        "### Preparing the data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iBjorjhhyUov"
      },
      "source": [
        "**Normalizing the data**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bm4FMO6jyUov"
      },
      "outputs": [],
      "source": [
        "mean = train_data.mean(axis=0)\n",
        "train_data -= mean\n",
        "std = train_data.std(axis=0)\n",
        "train_data /= std\n",
        "test_data -= mean\n",
        "test_data /= std"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xbmHgIm0yUov"
      },
      "source": [
        "### Building your model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mIjGE1BCyUow"
      },
      "source": [
        "**Model definition**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FAolWKxEyUow"
      },
      "outputs": [],
      "source": [
        "def build_model():\n",
        "    model = keras.Sequential([\n",
        "        layers.Dense(64, activation=\"relu\"),\n",
        "        layers.Dense(64, activation=\"relu\"),\n",
        "        layers.Dense(1)\n",
        "    ])\n",
        "    model.compile(optimizer=\"rmsprop\", loss=\"mse\", metrics=[\"mae\"])\n",
        "    return model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sO7BcBq-yUow"
      },
      "source": [
        "### Validating your approach using K-fold validation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NSKCAy94yUow"
      },
      "source": [
        "**K-fold validation**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "20pO79Z4yUow"
      },
      "outputs": [],
      "source": [
        "k = 4\n",
        "num_val_samples = len(train_data) // k\n",
        "num_epochs = 100\n",
        "all_scores = []\n",
        "for i in range(k):\n",
        "    print(f\"Processing fold #{i}\")\n",
        "    val_data = train_data[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    val_targets = train_targets[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    partial_train_data = np.concatenate(\n",
        "        [train_data[:i * num_val_samples],\n",
        "         train_data[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    partial_train_targets = np.concatenate(\n",
        "        [train_targets[:i * num_val_samples],\n",
        "         train_targets[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    model = build_model()\n",
        "    model.fit(partial_train_data, partial_train_targets,\n",
        "              epochs=num_epochs, batch_size=16, verbose=0)\n",
        "    val_mse, val_mae = model.evaluate(val_data, val_targets, verbose=0)\n",
        "    all_scores.append(val_mae)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_K6l5jG1yUow"
      },
      "outputs": [],
      "source": [
        "all_scores"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "AzA6s65uyUow"
      },
      "outputs": [],
      "source": [
        "np.mean(all_scores)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YEr95fKJyUox"
      },
      "source": [
        "**Saving the validation logs at each fold**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0BPq6yn2yUox"
      },
      "outputs": [],
      "source": [
        "num_epochs = 500\n",
        "all_mae_histories = []\n",
        "for i in range(k):\n",
        "    print(f\"Processing fold #{i}\")\n",
        "    val_data = train_data[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    val_targets = train_targets[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    partial_train_data = np.concatenate(\n",
        "        [train_data[:i * num_val_samples],\n",
        "         train_data[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    partial_train_targets = np.concatenate(\n",
        "        [train_targets[:i * num_val_samples],\n",
        "         train_targets[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    model = build_model()\n",
        "    history = model.fit(partial_train_data, partial_train_targets,\n",
        "                        validation_data=(val_data, val_targets),\n",
        "                        epochs=num_epochs, batch_size=16, verbose=0)\n",
        "    mae_history = history.history[\"val_mae\"]\n",
        "    all_mae_histories.append(mae_history)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qc086LqiyUox"
      },
      "source": [
        "**Building the history of successive mean K-fold validation scores**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JMGoY581yUox"
      },
      "outputs": [],
      "source": [
        "average_mae_history = [\n",
        "    np.mean([x[i] for x in all_mae_histories]) for i in range(num_epochs)]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dfIKJX1lyUox"
      },
      "source": [
        "**Plotting validation scores**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "W6e1b2TtyUox"
      },
      "outputs": [],
      "source": [
        "plt.plot(range(1, len(average_mae_history) + 1), average_mae_history)\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Validation MAE\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mH1pSIEGyUox"
      },
      "source": [
        "**Plotting validation scores, excluding the first 10 data points**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "P94Z7mdxyUoy"
      },
      "outputs": [],
      "source": [
        "truncated_mae_history = average_mae_history[10:]\n",
        "plt.plot(range(1, len(truncated_mae_history) + 1), truncated_mae_history)\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Validation MAE\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GWI_TCMZyUoy"
      },
      "source": [
        "**Training the final model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bbyFeSzLyUoy"
      },
      "outputs": [],
      "source": [
        "model = build_model()\n",
        "model.fit(train_data, train_targets,\n",
        "          epochs=130, batch_size=16, verbose=0)\n",
        "test_mse_score, test_mae_score = model.evaluate(test_data, test_targets)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "dJSeD7N4yUoy"
      },
      "outputs": [],
      "source": [
        "test_mae_score"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I4vmOvvDyUoz"
      },
      "source": [
        "### Generating predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XktK9qe9yUoz"
      },
      "outputs": [],
      "source": [
        "predictions = model.predict(test_data)\n",
        "predictions[0]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8yOKgLkwyUoz"
      },
      "source": [
        "### Wrapping up"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j2BOw4SKyUoz"
      },
      "source": [
        "## Summary"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "name": "chapter04_getting-started-with-neural-networks.i",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.0"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}