最近在看一篇Google的有关CTR预估的paper：Deep & Cross Network for Ad Click Predictions
整理记录一下参考的资料和自己的部分体会。

不错的参考资料

Deep and Cross Network原理及实现（有原理有代码）
Wide & Deep Learning for Recommender Systems Google的另一篇推荐系统的类似paper
Keras中文文档
[GitHub | DeepCTR] (https://github.com/shenweichen/DeepCTR)

遇到的问题

Deep and Cross Network原理及实现

• keras.layers.merge报错，需要修改为concatenate

from keras.layers import merge
input = merge(inp_embed, mode = 'concat')
output = merge([deep, cross], mode = 'concat')

参考文档：https://keras.io/layers/core/

from keras.layers import concatenate
input = concatenate(inp_embed, axis=-1)
output = concatenate([deep, cross], axis=-1)

• 画模型图需要安装额外的包

plot_model(model, to_file = 'model.png', show_shapes = True)

参考：https://stackoverflow.com/questions/40632486/dot-exe-not-found-in-path-pydot-on-python-windows-7

pip install django-extensions
pip install pyparsing
pip install graphviz
pip install pydot
conda install graphviz