Understanding Recommendation System Technology And Do Some Implementation Works.
- hongleizhang/RSPapers
- Chapter 9 about RS in MMDS
- A repo about RS papers
- A repo about RS system
- A repo about DL in RS
- Recommender Systems Handbook(2010)
- Recommender Systems Handbook 2nd(2018)
- RecSys 2017 summer slides: Deep Learning for Recommender Systems
- Recommender Systems Handbook 2nd(2015)
- ACM RecSys 2018 Late-Breaking Results Proceedings
- shenweichen/DeepCTR
- cheungdaven/DeepRec
- wubinzzu/NeuRec
- stasi009/Recommend-Estimators
- stasi009/NumpyWDL
- Implement Wide & Deep algorithm by using NumPy
- Content Based RS
- method
- Create vectors describing items, create vectors with the same components that describe the user's preferences.
- method
- Collaborative Filtering Based RS
- method
- The process of identifying similar users and recommending what similar users like.
- categories
- neighborhood methods
- latent factor models
- Matrix Factorization Based RS
- method
- Tree Based RS
- GBDT & Xgboost
- XGBoost + LR
- 兴趣树
- LR Related
- LR
- FM
- FFM
- Deep Learning Based RS
- Wide and Deep
- DeepFM
- DeepFFM
- Stream Based RS
- information filtering system
- Cold Start Problem
- 基于内容推荐
- 基于热点、时效推荐
- 增加额外信息，比如隐含信息(浏览/点击/购买)；增加兴趣时间动态信息，比如将用户隐向量看作是时间的函数等
- RS Explaination Problem
- The Long Tail Problem
- Scalability
- Accuracy
- RecSys
- NETFLIX PRIZE COMPETITION, 2006~2009
- Movie Lens
- Amazon Dataset
- Electronics (192,403 users, 63,001 goods, 801 categories and 1,689,188 samples.)
- Recommender Systems Datasets, by Julian McAuley, UCSD