cheungdaven / OpenCTR

A review and evaluation on CTR prediction models

Home Page:http://deepctr.readthedocs.io

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Documentation Status

OpenCTR

Click-through rate (CTR) prediction is an important task in many industrial applications such as online advertising, recommender systems, and sponsored search. OpenCTR builds an open-source library for benchmarking existing CTR prediction models.

Models

CTR prediction models currently available:

Model Conference Paper Available
LR WWW'07 Predicting Clicks: Estimating the Click-Through Rate for New Ads [Microsoft] ✔️
FM ICDM'10 Factorization Machines ✔️
FTRL KDD'13 Ad Click Prediction: a View from the Trenches [Google]
GBDT+LR ADKDD'14 Practical Lessons from Predicting Clicks on Ads at Facebook [Facebook]
CCPM CIKM'15 A Convolutional Click Prediction Model ✖️
FFM RecSys'16 Field-aware Factorization Machines for CTR Prediction [Criteo] ✔️
DNN RecSys'16 Deep Neural Networks for YouTube Recommendations [Google] ✔️
Wide&Deep DLRS'16 Wide & Deep Learning for Recommender Systems [Google] ✔️
FNN ECIR'16 Deep Learning over Multi-field Categorical Data: A Case Study on User Response Prediction [RayCloud] ✔️
PNN ICDM'16 Product-based Neural Networks for User Response Prediction ✔️
DeepFM IJCAI'17 DeepFM: A Factorization-Machine based Neural Network for CTR Prediction, [Huawei] ✔️
NFM SIGIR'17 Neural Factorization Machines for Sparse Predictive Analytics ✔️
AFM IJCAI'17 Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks ✔️
DCN ADKDD'17 Deep & Cross Network for Ad Click Predictions [Google] ✔️
FwFM WWW'18 Field-weighted Factorization Machines for Click-Through Rate Prediction in Display Advertising [Oath, TouchPal, LinkedIn, Ablibaba] ✖️
xDeepFM KDD'18 xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems [Microsoft] ✖️

Datasets

Dataset #Fields #Features #Instances #Train #Validation #Test Positive Ratio
Criteo 39 45,840,617 39,798,482 6,042,135 25.62%
Avazu 23 40,428,967 36,210,029 4,218,938 16.98%
iPinyou 14,758,717 12,237,087 2,521,630 0.0726%
Taobao
Alimama
Tencent
Bytedance

About

A review and evaluation on CTR prediction models

http://deepctr.readthedocs.io

License:MIT License