CTR ---Wide&Deep论文笔记，及tf2实现

记忆能力

泛化能力

提出动机

结构

交叉积变换

谷歌的实践

tf2实现

# coding:utf-8
# @Time: 2021/12/30 5:18 下午
# @File: ctr_WDL.py
'''
WDL
'''
'''
WDL
'''
import numpy as np
import tensorflow as tf
from tensorflow.keras import layers
from tensorflow.keras import Model
from tensorflow.keras import optimizers
from tensorflow.keras import metrics
from tensorflow.keras import regularizers
from sklearn.model_selection import train_test_split
import yaml

from tools import *


class WideDeepNet(Model):

    def __init__(self, spare_feature_columns, dense_feature_columns, hidden_units, output_dim, activation, droup_out, w_reg):
        super(WideDeepNet, self).__init__()
        self.dense_feature_columns = dense_feature_columns
        self.spare_feature_columns = spare_feature_columns
        self.w_reg = w_reg
        # embedding
        self.embedding_layer = {'embed_layer{}'.format(i): layers.Embedding(feat['vocabulary_size'], feat['embed_dim'])
                                for i, feat in enumerate(self.spare_feature_columns)}
        # deep
        self.DNN = tf.keras.Sequential()
        for hidden in hidden_units:
            self.DNN.add(layers.Dense(hidden))
            self.DNN.add(layers.BatchNormalization())
            self.DNN.add(layers.Activation(activation))
            self.DNN.add(layers.Dropout(droup_out))
        self.DNN.add(layers.Dense(output_dim, activation=None))

    def build(self, input_shape):
        self.w = self.add_weight(name='w', shape=(input_shape[-1], 1), initializer=tf.random_normal_initializer(),
                                 trainable=True, regularizer=regularizers.l2(self.w_reg))
        self.b = self.add_weight(name='b', shape=(1,), initializer=tf.zeros_initializer(), trainable=True)

    def call(self, inputs):
        # dense_inputs: 数值特征，13维
        # sparse_inputs： 类别特征，26维
        dense_inputs, sparse_inputs = inputs[:, :13], inputs[:, 13:]
        # wide
        wide_out = inputs @ self.w + self.b  # (batchsize, 1)
        # deep
        sparse_embed = tf.concat([self.embedding_layer['embed_layer{}'.format(i)](sparse_inputs[:, i]) for i in
                               range(sparse_inputs.shape[-1])], axis=-1)  # (batchsize, 26*embed_dim)
        dnn_input = tf.concat([sparse_embed, dense_inputs], axis=-1)  # (batchsize, 26*embed_dim + 13)
        deep_out = self.DNN(dnn_input)  # # (batchsize, 1)

        output = tf.sigmoid(wide_out + deep_out)
        return output


if __name__ == '__main__':
    with open('config.yaml', 'r') as f:
        config = yaml.Loader(f).get_data()

    data = pd.read_csv('../../data/criteo_sampled_data_OK.csv')
    data_X = data.iloc[:, 1:]
    data_y = data['label'].values

    # I1-I13：总共 13 列数值型特征
    # C1-C26：共有 26 列类别型特征
    dense_features = ['I' + str(i) for i in range(1, 14)]
    sparse_features = ['C' + str(i) for i in range(1, 27)]

    dense_feature_columns = [denseFeature(feat) for feat in dense_features]
    spare_feature_columns = [sparseFeature(feat, data_X[feat].nunique(), config['WDL']['embed_dim']) for feat in
                             sparse_features]

    tmp_X, test_X, tmp_y, test_y = train_test_split(data_X, data_y, test_size=0.2, random_state=42, stratify=data_y)
    train_X, val_X, train_y, val_y = train_test_split(tmp_X, tmp_y, test_size=0.1, random_state=42, stratify=tmp_y)

    model = WideDeepNet(spare_feature_columns=spare_feature_columns,
                        dense_feature_columns=dense_feature_columns,
                        hidden_units=config['WDL']['hidden_units'],
                        output_dim=config['WDL']['output_dim'],
                        activation=config['WDL']['activation'],
                        droup_out=config['WDL']['droup_out'],
                        w_reg=config['WDL']['w_reg'])

    adam = optimizers.Adam(lr=config['train']['adam_lr'], beta_1=0.95, beta_2=0.96,
                           decay=config['train']['adam_lr'] / config['train']['epochs'])
    model.compile(
        optimizer=adam,
        loss='binary_crossentropy',
        metrics=[metrics.AUC(), metrics.Precision(), metrics.Recall()]
    )
    model.fit(
        train_X.values, train_y,
        validation_data=(val_X.values, val_y),
        batch_size=config['train']['batch_size'],
        epochs=config['train']['epochs'],
        verbose=1,
    )

    model.summary()

Epoch 1/3
215/216 [============================>.] - ETA: 0s - loss: 362.3944 - auc: 0.5093 - precision: 0.2697 - recall: 0.2706Tensor("wide_deep_net/concat_1:0", shape=(2000, 117), dtype=float32)
216/216 [==============================] - 8s 37ms/step - loss: 360.7360 - auc: 0.5095 - precision: 0.2699 - recall: 0.2699 - val_loss: 9.5656 - val_auc: 0.5045 - val_precision: 0.2571 - val_recall: 0.8003
Epoch 2/3
216/216 [==============================] - 8s 36ms/step - loss: 8.1209 - auc: 0.5805 - precision: 0.3577 - recall: 0.3609 - val_loss: 4.0739 - val_auc: 0.5357 - val_precision: 0.6106 - val_recall: 0.0254
Epoch 3/3
216/216 [==============================] - 8s 36ms/step - loss: 2.9503 - auc: 0.6771 - precision: 0.4734 - recall: 0.4667 - val_loss: 5.3457 - val_auc: 0.5337 - val_precision: 0.6819 - val_recall: 0.0340