Skip to content

This is a kaggle challenge project called Display Advertising Challenge by CriteoLabs at 2014.这是2014年由CriteoLabs在kaggle上发起的广告点击率预估挑战项目。

License

Notifications You must be signed in to change notification settings

chengstone/kaggle_criteo_ctr_challenge-

Repository files navigation

kaggle_criteo_ctr_challenge-

This is a kaggle challenge project called Display Advertising Challenge by CriteoLabs at 2014. 这是2014年由CriteoLabs在kaggle上发起的广告点击率预估挑战项目。 使用TensorFlow1.0和Python 3.5开发。

Author chengstone

e-Mail [email protected]

代码详解请参见jupyter notebook和↓↓↓

知乎专栏:https://zhuanlan.zhihu.com/p/32500652

博客:http://blog.csdn.net/chengcheng1394/article/details/78940565

欢迎转发扩散 ^_^

本文使用GBDT、FM、FFM和神经网络构建了点击率预估模型。

网络模型

image

LogLoss曲线

image

验证集上的训练信息

  • 平均准确率
  • 平均损失
  • 平均Auc
  • 预测的平均点击率
  • 精确率、召回率、F1 Score等信息

image

更多内容请参考代码,Enjoy!

许可

Licensed under the MIT License with the 996ICU License.

About

This is a kaggle challenge project called Display Advertising Challenge by CriteoLabs at 2014.这是2014年由CriteoLabs在kaggle上发起的广告点击率预估挑战项目。

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published