算法介绍

左边deep network,右边FM,所以叫deepFM

1602764397990 - 推荐算法之: DeepFM及使用DeepCTR测试

包含两个部分:

  • Part1: FM(Factorization machines),因子分解机部分

1602764448337 - 推荐算法之: DeepFM及使用DeepCTR测试

在传统的一阶线性回归之上,加了一个二次项,可以表达两两特征的相互关系。

1602764632765 - 推荐算法之: DeepFM及使用DeepCTR测试

这里的公式可以简化,减少计算量,下图来至于网络。

1602765047458 - 推荐算法之: DeepFM及使用DeepCTR测试

  • Part2: Deep部分

deep部分是多层dnn网络。

算法实现

实现部分,用Keras实现一个DeepFM·清尘·《FM、FMM、DeepFM整理(pytorch)》

讲的比较清楚,这里引用keras实现来说明。

整体的网络结构:

1602764969628 - 推荐算法之: DeepFM及使用DeepCTR测试

特征编码

特征可以分为3类:

  • 连续型field,比如数字类型特征
  • 单值离散型特征,比如gender,可选为male、female
  • 多值离散型,比如tag,可以有多个

连续型field,可以拼接到一起,dense数据。

单值,多值field进行Onehot后,可见单值离散field对应的独热向量只有一位取1,而多值离散field对应的独热向量有多于一位取1,表示该field可以同时取多个特征值。

labelshop_scoregender=mgender=finterest=finterest=c
00.21011
10.80101

FM 部分

1602765456508 - 推荐算法之: DeepFM及使用DeepCTR测试

看公式:
1602764448337 - 推荐算法之: DeepFM及使用DeepCTR测试

先算 FM一次项:

  • 连续型field 可以用Dense(1)层实现
  • 单值离散型field 用Embedding(n,1), n是分类中值的个数
  • 多值离散型field可以同时取多个特征值,为了batch training,必须对样本进行补零padding。同样可以用Embedding实现,因为有多个Embedding,可以取下平均值。

1602765746903 - 推荐算法之: DeepFM及使用DeepCTR测试

然后计算FM二次项,这里理解比较费劲一点。

·清尘·《FM、FMM、DeepFM整理(pytorch)》 深入浅出的讲明白了这个过程,大家可以参见。

我们来看具体实现方面,这里的DeepFM模型CTR预估理论与实战
讲解更容易理解。

1602819038585 - 推荐算法之: DeepFM及使用DeepCTR测试

假设只有前面的C1和C2两个Category的特征,词典大小还是3和2。假设输入还是C1=2,C2=2(下标从1开始),则Embedding之后为V2=[e21,e22,e23,e24]和V5=[e51,e52,e53,e54]。

因为xi和xj同时不为零才需要计算,所以上面的公式里需要计算的只有i=2和j=5的情况。因此:

1602819112460 - 推荐算法之: DeepFM及使用DeepCTR测试

扩展到多个,比如C1,C2,C3,需要算内积

1602819135487 - 推荐算法之: DeepFM及使用DeepCTR测试

怎么用用矩阵乘法一次计算出来呢?我们可以看看这个

1602819176970 - 推荐算法之: DeepFM及使用DeepCTR测试

对应的代码就是

       square_of_sum = tf.square(reduce_sum(
			concated_embeds_value, axis=1, keep_dims=True))
		sum_of_square = reduce_sum(
			concated_embeds_value * concated_embeds_value, axis=1, keep_dims=True)
		cross_term = square_of_sum - sum_of_square
		cross_term = 0.5 * reduce_sum(cross_term, axis=2, keep_dims=False)

其中concated_embeds_value是拼接起来的embeds_value。

Deep部分

DNN比较简单,FM的输入和DNN的输入都是同一个group_embedding_dict。

使用movielens 来测试

下载ml-100k 数据集

wget http://files.grouplens.org/datasets/movielens/ml-100k.zip
unzip ml-100k.zip

安装相关软件包,sklearn,deepctr

导入包:

import pandas
import pandas as pd
import sklearn
from sklearn.metrics import log_loss, roc_auc_score
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from tensorflow.python.keras.preprocessing.sequence import pad_sequences
import tensorflow as tf
from tqdm import tqdm

from deepctr.models import DeepFM
from deepctr.feature_column import SparseFeat, VarLenSparseFeat, get_feature_names
import numpy as np

读取评分数据:

u_data = pd.read_csv("ml-100k/u.data", sep='\t', header=None)
u_data.columns = ['user_id', 'movie_id', 'rating', 'timestamp']

有评分的设置为1,随机采用未评分的

def neg_sample(u_data, neg_rate=1):
    # 全局随机采样
    item_ids = u_data['movie_id'].unique()
    print('start neg sample')
    neg_data = []
    # 负采样
    for user_id, hist in tqdm(u_data.groupby('user_id')):
        # 当前用户movie
        rated_movie_list = hist['movie_id'].tolist()
        candidate_set = list(set(item_ids) - set(rated_movie_list))
        neg_list_id = np.random.choice(candidate_set, size=len(rated_movie_list) * neg_rate, replace=True)
        for id in neg_list_id:
            neg_data.append([user_id, id, -1, 0])
    u_data_neg = pd.DataFrame(neg_data)
    u_data_neg.columns = ['user_id', 'movie_id', 'rating', 'timestamp']
    u_data = pandas.concat([u_data, u_data_neg])
    print('end neg sample')
    return u_data

读取item数据

u_item = pd.read_csv("ml-100k/u.item", sep='|', header=None, error_bad_lines=False)
    genres_columns = ['Action', 'Adventure',
                      'Animation',
                      'Children', 'Comedy', 'Crime', 'Documentary', 'Drama', 'Fantasy',
                      'Film_Noir', 'Horror', 'Musical', 'Mystery', 'Romance', 'Sci-Fi',
                      'Thriller', 'War', 'Western']

    u_item.columns = ['movie_id', 'title', 'release_date', 'video_date', 'url', 'unknown'] + genres_columns

处理genres并删除单独的genres列

     genres_list = []
    for index, row in u_item.iterrows():
        genres = []
        for item in genres_columns:
            if row[item]:
                genres.append(item)
        genres_list.append('|'.join(genres))

    u_item['genres'] = genres_list
    for item in genres_columns:
        del u_item[item]

读取用户信息:

  # user id | age | gender | occupation(职业) | zip code(邮编,地区)
    u_user = pd.read_csv("ml-100k/u.user", sep='|', header=None)
    u_user.columns = ['user_id', 'age', 'gender', 'occupation', 'zip']

join到一起:

 data = pandas.merge(u_data, u_item, on="movie_id", how='left')
 data = pandas.merge(data, u_user, on="user_id", how='left')
 data.to_csv('ml-100k/data.csv', index=False)

处理特征:

sparse_features = ["movie_id", "user_id",
                   "gender", "age", "occupation", "zip", ]

data[sparse_features] = data[sparse_features].astype(str)
target = ['rating']

# 评分
data['rating'] = [1 if int(x) >= 0 else 0 for x in data['rating']]

先特征编码:

for feat in sparse_features:
		lbe = LabelEncoder()
		data[feat] = lbe.fit_transform(data[feat])

处理genres特征,一个movie有多个genres,先拆分,然后编码为数字,注意是从1开始;由于每个movie的genres长度不一样,可以计算最大长度,位数不足的后面补零(pad_sequences,在post补0)

	 def split(x):
			key_ans = x.split('|')
			for key in key_ans:
				if key not in key2index:
					# Notice : input value 0 is a special "padding",so we do not use 0 to encode valid feature for sequence input
					key2index[key] = len(key2index) + 1
			return list(map(lambda x: key2index[x], key_ans))


		key2index = {}
		genres_list = list(map(split, data['genres'].values))
		genres_length = np.array(list(map(len, genres_list)))
		max_len = max(genres_length)
		# Notice : padding=`post`
		genres_list = pad_sequences(genres_list, maxlen=max_len, padding='post', )

构建deepctr的特征列,主要分为两类特征,一是定长的SparseFeat,稀疏的类别特征,二是可变长度的VarLenSparseFeat,像genres这样的包含多个的。

	   fixlen_feature_columns = [SparseFeat(feat, data[feat].nunique(), embedding_dim=4)
								  for feat in sparse_features]

		use_weighted_sequence = False
		if use_weighted_sequence:
			varlen_feature_columns = [VarLenSparseFeat(SparseFeat('genres', vocabulary_size=len(
				key2index) + 1, embedding_dim=4), maxlen=max_len, combiner='mean',
													   weight_name='genres_weight')]  # Notice : value 0 is for padding for sequence input feature
		else:
			varlen_feature_columns = [VarLenSparseFeat(SparseFeat('genres', vocabulary_size=len(
				key2index) + 1, embedding_dim=4), maxlen=max_len, combiner='mean',
													   weight_name=None)]  # Notice : value 0 is for padding for sequence input feature

		linear_feature_columns = fixlen_feature_columns + varlen_feature_columns
		dnn_feature_columns = fixlen_feature_columns + varlen_feature_columns

		feature_names = get_feature_names(linear_feature_columns + dnn_feature_columns)

封装训练数据,先shuffle(乱排)数据,然后生成dict input数据。

data = sklearn.utils.shuffle(data)
train_model_input = {name: data[name] for name in sparse_features}  #
train_model_input["genres"] = genres_list

构建DeepFM模型,由于目标值是0,1,因此采用binary,损失函数用binary_crossentropy

model = DeepFM(linear_feature_columns, dnn_feature_columns, task='binary')

model.compile(optimizer=tf.keras.optimizers.Adam(), loss='binary_crossentropy',
              metrics=['AUC', 'Precision', 'Recall'])
model.summary()

训练模型:

model.fit(train_model_input, data[target].values,
						batch_size=256, epochs=20, verbose=2,
						validation_split=0.2
				)

开始训练:

Epoch 1/20
625/625 - 3s - loss: 0.5081 - auc: 0.8279 - precision: 0.7419 - recall: 0.7695 - val_loss: 0.4745 - val_auc: 0.8513 - val_precision: 0.7563 - val_recall: 0.7936
Epoch 2/20
625/625 - 2s - loss: 0.4695 - auc: 0.8538 - precision: 0.7494 - recall: 0.8105 - val_loss: 0.4708 - val_auc: 0.8539 - val_precision: 0.7498 - val_recall: 0.8127
Epoch 3/20
625/625 - 2s - loss: 0.4652 - auc: 0.8564 - precision: 0.7513 - recall: 0.8139 - val_loss: 0.4704 - val_auc: 0.8545 - val_precision: 0.7561 - val_recall: 0.8017
Epoch 4/20
625/625 - 2s - loss: 0.4624 - auc: 0.8579 - precision: 0.7516 - recall: 0.8146 - val_loss: 0.4724 - val_auc: 0.8542 - val_precision: 0.7296 - val_recall: 0.8526
Epoch 5/20
625/625 - 2s - loss: 0.4607 - auc: 0.8590 - precision: 0.7521 - recall: 0.8173 - val_loss: 0.4699 - val_auc: 0.8550 - val_precision: 0.7511 - val_recall: 0.8141
Epoch 6/20
625/625 - 2s - loss: 0.4588 - auc: 0.8602 - precision: 0.7545 - recall: 0.8165 - val_loss: 0.4717 - val_auc: 0.8542 - val_precision: 0.7421 - val_recall: 0.8265
Epoch 7/20
625/625 - 2s - loss: 0.4574 - auc: 0.8610 - precision: 0.7535 - recall: 0.8192 - val_loss: 0.4722 - val_auc: 0.8547 - val_precision: 0.7549 - val_recall: 0.8023
Epoch 8/20
625/625 - 2s - loss: 0.4561 - auc: 0.8619 - precision: 0.7543 - recall: 0.8201 - val_loss: 0.4717 - val_auc: 0.8548 - val_precision: 0.7480 - val_recall: 0.8185
Epoch 9/20
625/625 - 2s - loss: 0.4531 - auc: 0.8643 - precision: 0.7573 - recall: 0.8210 - val_loss: 0.4696 - val_auc: 0.8583 - val_precision: 0.7598 - val_recall: 0.8103
Epoch 10/20
625/625 - 2s - loss: 0.4355 - auc: 0.8768 - precision: 0.7787 - recall: 0.8166 - val_loss: 0.4435 - val_auc: 0.8769 - val_precision: 0.7756 - val_recall: 0.8293
Epoch 11/20
625/625 - 2s - loss: 0.4093 - auc: 0.8923 - precision: 0.7915 - recall: 0.8373 - val_loss: 0.4301 - val_auc: 0.8840 - val_precision: 0.7806 - val_recall: 0.8390
Epoch 12/20
625/625 - 2s - loss: 0.3970 - auc: 0.8988 - precision: 0.7953 - recall: 0.8497 - val_loss: 0.4286 - val_auc: 0.8867 - val_precision: 0.7903 - val_recall: 0.8299
Epoch 13/20
625/625 - 2s - loss: 0.3896 - auc: 0.9029 - precision: 0.8001 - recall: 0.8542 - val_loss: 0.4253 - val_auc: 0.8888 - val_precision: 0.7913 - val_recall: 0.8322
Epoch 14/20
625/625 - 2s - loss: 0.3825 - auc: 0.9067 - precision: 0.8038 - recall: 0.8584 - val_loss: 0.4205 - val_auc: 0.8917 - val_precision: 0.7885 - val_recall: 0.8506
Epoch 15/20
625/625 - 2s - loss: 0.3755 - auc: 0.9102 - precision: 0.8074 - recall: 0.8624 - val_loss: 0.4204 - val_auc: 0.8940 - val_precision: 0.7868 - val_recall: 0.8607
Epoch 16/20
625/625 - 2s - loss: 0.3687 - auc: 0.9136 - precision: 0.8117 - recall: 0.8653 - val_loss: 0.4176 - val_auc: 0.8956 - val_precision: 0.8097 - val_recall: 0.8236
Epoch 17/20
625/625 - 2s - loss: 0.3617 - auc: 0.9170 - precision: 0.8155 - recall: 0.8682 - val_loss: 0.4166 - val_auc: 0.8966 - val_precision: 0.8056 - val_recall: 0.8354
Epoch 18/20
625/625 - 2s - loss: 0.3553 - auc: 0.9201 - precision: 0.8188 - recall: 0.8716 - val_loss: 0.4168 - val_auc: 0.8977 - val_precision: 0.7996 - val_recall: 0.8492
Epoch 19/20
625/625 - 2s - loss: 0.3497 - auc: 0.9227 - precision: 0.8214 - recall: 0.8741 - val_loss: 0.4187 - val_auc: 0.8973 - val_precision: 0.8079 - val_recall: 0.8358
Epoch 20/20
625/625 - 2s - loss: 0.3451 - auc: 0.9248 - precision: 0.8244 - recall: 0.8753 - val_loss: 0.4210 - val_auc: 0.8982 - val_precision: 0.7945 - val_recall: 0.8617

最后我们测试下数据:

 pred_ans = model.predict(train_model_input, batch_size=256)
 count = 0
    for (i, j) in zip(pred_ans, data['rating'].values):
        print(i, j)
        count += 1
        if count > 10:
            break

输出如下:

[0.20468083] 0
[0.1988303] 0
[7.7236204e-05] 0
[0.9439401] 1
[0.76648283] 0
[0.80082995] 1
[0.7689271] 0
[0.8515004] 1
[0.93311656] 1
[0.40019292] 0
[0.60735244] 0

参考

作者:Jadepeng
出处:jqpeng的技术记事本–http://www.cnblogs.com/xiaoqi
您的支持是对博主最大的鼓励,感谢您的认真阅读。
本文版权归作者所有,欢迎转载,但未经作者同意必须保留此段声明,且在文章页面明显位置给出原文连接,否则保留追究法律责任的权利。

推荐算法之: DeepFM及使用DeepCTR测试

推荐算法之: DeepFM及使用DeepCTR测试插图

发表评论