深度推荐模型——DIN [KDD 18][Alibaba]

重磅推荐专栏： 《Transformers自然语言处理系列教程》
手把手带你深入实践Transformers，轻松构建属于自己的NLP智能应用！

import tensorflow as tf
from tensorflow import keras
from utils import *

EPOCH = 10
BATCH_SIZE = 32
VEC_DIM = 10
DROPOUT_RATE = 0.5
HEAD_NUM = 4
HIDE_SIZE = 128
LAYER_NUM = 3
DNN_LAYERS = [200, 80]
data, max_user_id, max_item_id = load_data()
# 行为特征个数
BEHAVIOR_FEAT_NUM = 32
K = tf.keras.backend


def run():
    train_user_id_data, train_now_item_id_data, train_item_ids_data, train_rating_ids_data, train_y_data,\
    test_user_id_data, test_now_item_id_data, test_item_ids_data, test_rating_ids_data, test_y_data = get_all_data(data)

    user_id = keras.Input((1,))
    now_item_id = keras.Input((1,))
    items_ids = keras.Input((BEHAVIOR_FEAT_NUM,))
    ratings_ids = keras.Input((BEHAVIOR_FEAT_NUM,))

    usr_emb = keras.layers.Embedding(max_user_id + 1, VEC_DIM, input_length=1)(user_id)  # [-1,1,vec_dim]
    usr_emb = keras.layers.Flatten()(usr_emb)  # [-1,vec_dim]
    now_item_emb = keras.layers.Embedding(max_item_id + 1, VEC_DIM, input_length=1)(now_item_id)  # [-1,1,vec_dim]
    now_item_emb = keras.layers.Flatten()(now_item_emb)  # [-1,vec_dim]
    items_emb = keras.layers.Embedding(max_item_id + 1, VEC_DIM, input_length=BEHAVIOR_FEAT_NUM)(
        items_ids)  # [-1,BEA_FEAT_NUM,vec_dim]
    ratings_emb = keras.layers.Embedding(6, VEC_DIM, input_length=BEHAVIOR_FEAT_NUM)(
        ratings_ids)  # [-1,BEA_FEAT_NUM,vec_dim]
    behavior_emb = keras.layers.concatenate([items_emb, ratings_emb])  # [-1,BEA_FEAT_NUM, 2 * vec_dim]
    behavior_emb = tf.reduce_sum(behavior_emb, axis=1)  # [-1, 2 * vec_dim]

    deep = keras.layers.concatenate([usr_emb, now_item_emb, behavior_emb])

    for units in DNN_LAYERS:
        deep = keras.layers.Dense(units)(deep)
        deep = keras.layers.PReLU()(deep)
        deep = keras.layers.Dropout(DROPOUT_RATE)(deep)
    outputs = keras.layers.Dense(1, activation='sigmoid')(deep)

    model = keras.Model(inputs=[user_id, now_item_id, items_ids, ratings_ids], outputs=outputs)
    model.compile(loss='binary_crossentropy', optimizer=tf.train.AdamOptimizer(0.001), metrics=[keras.metrics.AUC()])
    tbCallBack = keras.callbacks.TensorBoard(log_dir='./logs',
                                             histogram_freq=0,
                                             write_graph=True,
                                             write_grads=True,
                                             write_images=True,
                                             embeddings_freq=0,
                                             embeddings_layer_names=None,
                                             embeddings_metadata=None)

    model.fit([train_user_id_data, train_now_item_id_data, train_item_ids_data, train_rating_ids_data], train_y_data,
              batch_size=BATCH_SIZE, epochs=EPOCH, verbose=2,
              validation_data=(
              [test_user_id_data, test_now_item_id_data, test_item_ids_data, test_rating_ids_data], test_y_data),
              callbacks=[tbCallBack])


run()