该仓库是自己学习CTR预估模型过程中,利用Keras
实现的经典模型的代码。随着学习的深入,还在不断更新中......
很多模型作者都公布了源代码,可以在原论文中找到。
- 模型名 (论文title)
- LR (Logistic Regression)
- DNN (Deep Neural Networks)
- LR-GBDT (Practical Lessons from Predicting Clicks on Ads at Facebook)
- PNN (Product-based Neural Networks for User Response Prediction)
- NFM (Neural Factorization Machine for Sparse Predictive Analytics)
- AFM (Attentional Factorization Machines-Learning the Weight of Feature Interactions via Attention Networks)
- DCN (Deep & Cross Network for Ad Click Predictions)
- DeepFM (A Factorization-Machine based Neural Network for CTR Prediction)
- xDeepFM (Combining Explicit and Implicit Feature Interactions for Recommender Systems)