神级程序员带来的：基于Python和Tensorflow的电影推

userIdmovieIdratingtimestamp

9999967162682.51065579370

10000067162694.01065149201

10000167163654.01070940363

10000267163852.51070979663

10000367165653.51074784724

movieIdtitlegenres

9120162672Mohenjo Daro (2016)Adventure|Drama|Romance

9121163056Shin Godzilla (2016)Action|Adventure|Fantasy|Sci-Fi

9122163949The Beatles: Eight Days a Week - The Touring Y...Documentary

9123164977The Gay Desperado (1936)Comedy

9124164979Women of '69, UnboxedDocumentary

movieIdtitlegenresmovieRow

9120162672Mohenjo Daro (2016)Adventure|Drama|Romance9120

9121163056Shin Godzilla (2016)Action|Adventure|Fantasy|Sci-Fi9121

9122163949The Beatles: Eight Days a Week - The Touring Y...Documentary9122

9123164977The Gay Desperado (1936)Comedy9123

9124164979Women of '69, UnboxedDocumentary9124

movieRowmovieIdtitle

91209120162672Mohenjo Daro (2016)

91219121163056Shin Godzilla (2016)

91229122163949The Beatles: Eight Days a Week - The Touring Y...

91239123164977The Gay Desperado (1936)

91249124164979Women of '69, Unboxed

userIdmovieRowrating

01302.5

17303.0

231304.0

332304.0

436303.0

第三步：构建模型

loss =1/2* tf.reduce_sum(((tf.matmul(X_parameters, Theta_parameters, transpose_b = True) - rating_norm) * record) **2) +1/2* (tf.reduce_sum(X_parameters **2) + tf.reduce_sum(Theta_parameters **2))#基于内容的推荐算法模型

函数解释：

reduce_sum() 就是求和，reduce_sum( input_tensor, axis=None, keep_dims=False, name=None, reduction_indices=None)

reduce_sum() 参数解释：

查看训练结果

在终端输入 tensorboard --logir=./

第五步：评估模型

4037.9002717628305

第六步：构建完整的电影推荐系统

本文章内容来自慕课网，仅供学习，如有侵权，联系删除

欢迎大家关注我的博客：https://home.cnblogs.com/u/sm123456/

欢迎大家加入千人交流学习群：125240963