DeepFM原理以及DeepCTR代码实现

87 阅读 0 评论 58 点赞

我是靠谱客的博主默默冰淇淋，这篇文章主要介绍DeepFM原理以及DeepCTR代码实现，现在分享给大家，希望可以做个参考。

DeepFM算法：

论文： A Factorization-Machine based Neural Network for CTR Prediction，2017
https://arxiv.org/abs/1703.04247

FM（因子分解机，之前的博文有讲过）可以做特征组合，但是计算量大，一般只考虑2阶特征组合
如何既考虑低阶（1阶+2阶），又能考虑到高阶特征 => DeepFM=FM+DNN
设计了一种end-to-end的模型结构 => 无须特征工程
在各种benchmark和工程中效果好
Criteo点击率预测, 4500万用户点击记录，90%样本用于训练，10%用于测试
Company*游戏中心，10亿记录，连续7天用户点击记录用于训练，之后1天用于测试

复制代码

import pandas as pd
from sklearn.metrics import mean_squared_error
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from deepctr.models import DeepFM
from deepctr.feature_column import SparseFeat,get_feature_names

#数据加载
data = pd.read_csv("movielens_sample.txt")
sparse_features = ["movie_id", "user_id", "gender", "age", "occupation", "zip"]
target = ['rating']

# 对特征标签进行编码
for feature in sparse_features:
    lbe = LabelEncoder()
    data[feature] = lbe.fit_transform(data[feature])
# 计算每个特征中的 不同特征值的个数
fixlen_feature_columns = [SparseFeat(feature, data[feature].nunique()) for feature in sparse_features]
linear_feature_columns = fixlen_feature_columns
dnn_feature_columns = fixlen_feature_columns
feature_names = get_feature_names(linear_feature_columns + dnn_feature_columns)

# 将数据集切分成训练集和测试集
train, test = train_test_split(data, test_size=0.2)
train_model_input = {name:train[name].values for name in feature_names}
test_model_input = {name:test[name].values for name in feature_names}

# 使用DeepFM进行训练
model = DeepFM(linear_feature_columns, dnn_feature_columns, task='regression')
model.compile("adam", "mse", metrics=['mse'], )
history = model.fit(train_model_input, train[target].values, batch_size=256, epochs=1, verbose=True, validation_split=0.2, )
# 使用DeepFM进行预测
pred_ans = model.predict(test_model_input, batch_size=256)
# 输出RMSE或MSE
mse = round(mean_squared_error(test[target].values, pred_ans), 4)
rmse = mse ** 0.5
print("test00 RMSE", rmse)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
import pandas as pd
from sklearn.metrics import mean_squared_error
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from deepctr.models import DeepFM
from deepctr.feature_column import SparseFeat,get_feature_names

#数据加载
data = pd.read_csv("movielens_sample.txt")
sparse_features = ["movie_id", "user_id", "gender", "age", "occupation", "zip"]
target = ['rating']

# 对特征标签进行编码
for feature in sparse_features:
    lbe = LabelEncoder()
    data[feature] = lbe.fit_transform(data[feature])
# 计算每个特征中的 不同特征值的个数
fixlen_feature_columns = [SparseFeat(feature, data[feature].nunique()) for feature in sparse_features]
linear_feature_columns = fixlen_feature_columns
dnn_feature_columns = fixlen_feature_columns
feature_names = get_feature_names(linear_feature_columns + dnn_feature_columns)

# 将数据集切分成训练集和测试集
train, test = train_test_split(data, test_size=0.2)
train_model_input = {name:train[name].values for name in feature_names}
test_model_input = {name:test[name].values for name in feature_names}

# 使用DeepFM进行训练
model = DeepFM(linear_feature_columns, dnn_feature_columns, task='regression')
model.compile("adam", "mse", metrics=['mse'], )
history = model.fit(train_model_input, train[target].values, batch_size=256, epochs=1, verbose=True, validation_split=0.2, )
# 使用DeepFM进行预测
pred_ans = model.predict(test_model_input, batch_size=256)
# 输出RMSE或MSE
mse = round(mean_squared_error(test[target].values, pred_ans), 4)
rmse = mse ** 0.5
print("test00 RMSE", rmse)