FM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "53de0a97-5762-4534-bc15-bca11bcf8530",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "import tensorflow.keras as keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e8db25cc-5b29-4c04-9baa-bf512b6c9ba1",
   "metadata": {},
   "outputs": [],
   "source": [
    "TARGET_FEATURE_NAME = \"label\"\n",
    "TARGET_FEATURE_LABELS = [\"0\", \"1\"]\n",
    "NUM_CLASSES = len(TARGET_FEATURE_LABELS)\n",
    "\n",
    "# Numeric Feature\n",
    "NUMERIC_FEATURE_NAMES = [\n",
    "    'movieAvgRating',\n",
    "    'movieRatingStddev',\n",
    "    'movieRatingCount',\n",
    "    'userAvgRating',\n",
    "    'userRatingStddev',\n",
    "    'userRatingCount',\n",
    "    'releaseYear',\n",
    "]\n",
    "\n",
    "# ID Feature\n",
    "ID_FEATURE_NAMES = ['movieId', 'userId']\n",
    "\n",
    "# Categorical Feature\n",
    "genre_vocab = ['Film-Noir', 'Action', 'Adventure', \n",
    "               'Horror', 'Romance', 'War', \n",
    "               'Comedy', 'Western', 'Documentary',\n",
    "               'Sci-Fi', 'Drama', 'Thriller',\n",
    "               'Crime', 'Fantasy', 'Animation', \n",
    "               'IMAX', 'Mystery', 'Children', 'Musical']\n",
    "\n",
    "CATEGORICAL_FEATURES_WITH_VOCABULARY = {\n",
    "    'userGenre1': genre_vocab,\n",
    "    'userGenre2': genre_vocab,\n",
    "    'userGenre3': genre_vocab,\n",
    "    'userGenre4': genre_vocab,\n",
    "    'userGenre5': genre_vocab,\n",
    "    'movieGenre1': genre_vocab,\n",
    "    'movieGenre2': genre_vocab,\n",
    "    'movieGenre3': genre_vocab\n",
    "}\n",
    "\n",
    "CATEGORICAL_FEATURE_NAMES = list(CATEGORICAL_FEATURES_WITH_VOCABULARY.keys())\n",
    "FEATURE_NAMES = NUMERIC_FEATURE_NAMES + CATEGORICAL_FEATURE_NAMES + ID_FEATURE_NAMES\n",
    "\n",
    "# 读取CSV时的缺失值填充\n",
    "COLUMN_DEFAULTS = [\n",
    "    [0.0] if feature_name in NUMERIC_FEATURE_NAMES + [TARGET_FEATURE_NAME] else [\"NA\"]\n",
    "    for feature_name in FEATURE_NAMES\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "399153f0-4d2b-40ab-b0a8-7e30d81f77d7",
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_model_inputs():\n",
    "    inputs = {}\n",
    "    for feature_name in FEATURE_NAMES:\n",
    "        if feature_name in NUMERIC_FEATURE_NAMES:\n",
    "            inputs[feature_name] = keras.layers.Input(\n",
    "                name=feature_name, shape=(), dtype=tf.float32\n",
    "            )\n",
    "        if feature_name in CATEGORICAL_FEATURE_NAMES:\n",
    "            inputs[feature_name] = keras.layers.Input(\n",
    "                name=feature_name, shape=(), dtype=tf.string\n",
    "            )\n",
    "        if feature_name in ID_FEATURE_NAMES:\n",
    "            inputs[feature_name] = keras.layers.Input(\n",
    "                name=feature_name, shape=(), dtype=tf.string\n",
    "            )\n",
    "    return inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "6b8431aa-2459-48f8-bd2d-b6bebee4e44e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def encode_fm_inputs(inputs, embedding_dims=10):\n",
    "    encoded_cols = []\n",
    "    for feature_name in inputs:\n",
    "        if feature_name in CATEGORICAL_FEATURE_NAMES:\n",
    "            cat_col = tf.feature_column.categorical_column_with_hash_bucket(\n",
    "                key=feature_name, hash_bucket_size=10)\n",
    "            emb_col = tf.feature_column.embedding_column(cat_col, embedding_dims)\n",
    "            \n",
    "        elif feature_name in ID_FEATURE_NAMES:\n",
    "            cat_col = tf.feature_column.categorical_column_with_hash_bucket(\n",
    "                key=feature_name, hash_bucket_size=1000)\n",
    "            emb_col = tf.feature_column.embedding_column(cat_col, embedding_dims)\n",
    "            \n",
    "        elif feature_name in NUMERIC_FEATURE_NAMES:\n",
    "            num = tf.feature_column.numeric_column(key=feature_name)\n",
    "            bucket = tf.feature_column.bucketized_column(num, boundaries = [0,5,10])\n",
    "            emb_col = tf.feature_column.embedding_column(bucket, embedding_dims)\n",
    "#             emb_col = tf.feature_column.numeric_column(feature_name)\n",
    "            \n",
    "        encoded_cols.append(emb_col)\n",
    "    return encoded_cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "c8864108-2c4a-4869-8b21-594752c83ecb",
   "metadata": {},
   "outputs": [],
   "source": [
    "class FM_Layer(keras.layers.Layer):\n",
    "    def __init__(self, feature_num, k=10, w_reg=1e-6):\n",
    "        \"\"\"\n",
    "        Factorization Machines\n",
    "        :param feature_columns: A list. sparse column feature information.\n",
    "        :param k: the latent vector\n",
    "        :param w_reg: the regularization coefficient of parameter w\n",
    "        :param v_reg: the regularization coefficient of parameter v\n",
    "        \"\"\"\n",
    "        super(FM_Layer, self).__init__()\n",
    "        self.k = k\n",
    "        self.w_reg = w_reg\n",
    "        self.feature_num = feature_num\n",
    "        \n",
    "    def build(self, input_shape):\n",
    "        print(input_shape)\n",
    "        self.bias = self.add_weight(name='bias', shape=(1,),\n",
    "                                  initializer=tf.zeros_initializer(),\n",
    "                                  trainable=True)\n",
    "        self.W = self.add_weight(name='W', shape=(input_shape[1], 1),\n",
    "                                 initializer=tf.random_normal_initializer(),\n",
    "                                 regularizer=keras.regularizers.l2(self.w_reg),\n",
    "                                 trainable=True)\n",
    "\n",
    "    def call(self, inputs, **kwargs):\n",
    "        # linear_part\n",
    "        linear_part = self.bias + tf.matmul(inputs, self.W)  # (batch_size, 1)\n",
    "        print(linear_part.shape)\n",
    "        \n",
    "        # interaction_part\n",
    "        fm_inputs = tf.reshape(inputs, (-1, self.feature_num, self.k))\n",
    "        print(fm_inputs.shape)\n",
    "        square_sum = tf.square(tf.reduce_sum(fm_inputs, axis=1))  # (batch_size, embed_dim)\n",
    "        sum_square = tf.reduce_sum(tf.square(fm_inputs), axis=1)  # (batch_size, embed_dim)\n",
    "        interaction_part = 0.5 * tf.reduce_sum(square_sum - sum_square, axis=1)  # (batch_size,)\n",
    "        interaction_part = tf.expand_dims(interaction_part, 1) # (batch_size, 1)\n",
    "        print(interaction_part.shape)\n",
    "        \n",
    "        # outputs\n",
    "        outputs = linear_part + interaction_part\n",
    "        return outputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6f7a7d9e-8148-46d3-ac73-0e912d69b81b",
   "metadata": {},
   "outputs": [],
   "source": [
    "learning_rate = 0.001\n",
    "dropout_rate = 0.1\n",
    "batch_size = 12\n",
    "num_epochs = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7db21327-bbae-4e64-b8e4-92fc508a5b57",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(None, 170)\n",
      "(None, 1)\n",
      "(None, 17, 10)\n",
      "(None, 1)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def create_FM_model():\n",
    "    embedding_dims = 10\n",
    "    inputs = create_model_inputs()\n",
    "    encoded_cols = encode_fm_inputs(inputs=inputs, embedding_dims=embedding_dims)\n",
    "    fm_inputs = tf.keras.layers.DenseFeatures(encoded_cols)(inputs)\n",
    "    fm_outputs = FM_Layer(feature_num=17, k=embedding_dims)(fm_inputs)\n",
    "    outputs = keras.layers.Dense(1, activation='sigmoid')(fm_outputs)\n",
    "    model = tf.keras.Model(inputs, outputs)\n",
    "    return model\n",
    "\n",
    "FM_model = create_FM_model()\n",
    "keras.utils.plot_model(FM_model, show_shapes=True, rankdir=\"TB\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "bd088c89-d38e-4a9a-82ce-4ae4bc34bf3b",
   "metadata": {},
   "outputs": [],
   "source": [
    "training_samples_file_path = './sampledata/trainingSamples.csv'\n",
    "test_samples_file_path = './sampledata/testSamples.csv'\n",
    "\n",
    "def get_dataset(file_path, batch_size):\n",
    "    dataset = tf.data.experimental.make_csv_dataset(\n",
    "        file_path,\n",
    "        batch_size=batch_size,\n",
    "        label_name='label',\n",
    "        na_value=\"0\",\n",
    "        num_epochs=1,\n",
    "        ignore_errors=True)\n",
    "    return dataset.cache()\n",
    "\n",
    "batch_size = 128\n",
    "# split as test dataset and training dataset\n",
    "train_dataset = get_dataset(training_samples_file_path, batch_size)\n",
    "test_dataset = get_dataset(test_samples_file_path, batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "01f22385-1529-414f-8b13-6c2930656754",
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_experiment(model): \n",
    "    model.compile(\n",
    "        optimizer=keras.optimizers.Adam(),\n",
    "        loss=keras.losses.BinaryCrossentropy(),\n",
    "        metrics=['accuracy', tf.keras.metrics.AUC(curve='ROC'), tf.keras.metrics.AUC(curve='PR')],\n",
    "    )\n",
    "    print(\"Start training the model...\")\n",
    "    history = model.fit(train_dataset, epochs=num_epochs)\n",
    "    print(\"Model training finished\")\n",
    "    _, accuracy, _, _ = model.evaluate(test_dataset, verbose=0)\n",
    "    print(f\"Test accuracy: {round(accuracy * 100, 2)}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7b7b36ae-8763-48a1-8c6e-d5d56bca1a39",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start training the model...\n",
      "Epoch 1/3\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Conda\\envs\\studio\\lib\\site-packages\\tensorflow\\python\\keras\\engine\\functional.py:595: UserWarning: Input dict contained keys ['rating', 'timestamp', 'userRatedMovie1', 'userRatedMovie2', 'userRatedMovie3', 'userRatedMovie4', 'userRatedMovie5', 'userAvgReleaseYear', 'userReleaseYearStddev'] which did not match any model input. They will be ignored by the model.\n",
      "  [n for n in tensors.keys() if n not in ref_input_names])\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(None, 1)\n",
      "(None, 17, 10)\n",
      "(None, 1)\n",
      "(None, 1)\n",
      "(None, 17, 10)\n",
      "(None, 1)\n",
      "694/694 [==============================] - 8s 5ms/step - loss: 0.8484 - accuracy: 0.5302 - auc: 0.5280 - auc_1: 0.5860\n",
      "Epoch 2/3\n",
      "694/694 [==============================] - 4s 5ms/step - loss: 0.6454 - accuracy: 0.6182 - auc: 0.6587 - auc_1: 0.7082\n",
      "Epoch 3/3\n",
      "694/694 [==============================] - 4s 6ms/step - loss: 0.6150 - accuracy: 0.6575 - auc: 0.7112 - auc_1: 0.7505\n",
      "Model training finished\n",
      "(None, 1)\n",
      "(None, 17, 10)\n",
      "(None, 1)\n",
      "Test accuracy: 65.62%\n"
     ]
    }
   ],
   "source": [
    "run_experiment(FM_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "99610384-d062-4f7a-ad09-ee515c8469d5",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:studio]",
   "language": "python",
   "name": "conda-env-studio-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}