Recommend system dry goods summary

This article was first published on the public account: Code Nongxiu Refining Factory

preface

Recommendation system is a very hot research direction, in the industry and academia have been widely concerned. Hope that through this article, summarizes some about the meeting, well-known scholars in the fields of recommendation system, and do research of commonly used data sets, code library, etc., would be to dabble in sorting and summary in the field of recommender systems, and expected to help the introductory to recommendation system of children’s shoes to provide a reference, hope to be able to fit recommendation system as soon as possible, Further better and faster in-depth scientific research or engineering.

I. Related meetings

As for the field of recommendation system, there are not many conferences directly related to it. However, as recommendation system involves data mining, machine learning and other aspects of knowledge, and as one of the important applications of data mining and machine learning, recommendation system also belongs to the category of artificial intelligence if it moves closer to a larger field. So many referrals are also looking at conferences on data mining, machine learning and artificial intelligence. Therefore, if we want to focus on the cutting edge of recommendation systems, we need to focus not only on the annual recommendation systems conference, but also on other referral-related conferences.

1. Meetings directly related to the recommendation system

RecSys -The ACM Conference Series on Recommender Systems.

2. Data mining related meetings

SIGKDD – The ACM SIGKDD Conference on Knowledge Discovery and Data Mining.

WSDM – The International Conference on Web Search and Data Mining.

ICDM – The IEEE International Conference on Data Mining.

SDM -TheSIAM International Conference on Data Mining.

ECML-PKDD – The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases

3. Machine learning related conferences

ICML – The International Conference on Machine Learning.

NIPS – The Conference on Neural Information Processing Systems

4. Information retrieval related meetings

SIGIR – The ACM International Conference on Research and Development in Information Retrieval

5. Database related meetings

CIKM – The ACM International Conference on Information and Knowledge Management.

6. Web-related meetings

WWW – The International World Wide Web Conference.

7. Meetings on ARTIFICIAL intelligence

AAAI – The National Conference of the American Association for Artificial Intelligence.

IJCAI – The International Joint Conference on Artificial Intelligence.

ECAI -European Conference on Artificial Intelligence

UAI – The Conference on Uncertainty in Artificial Intelligence

Ii. Relevant scholars

(Note: in no particular order…)

Yehuda Koren

Personal HomePage: Koren’s HomePage

Major contribution: The winner of the Netflix Prize is a god in the field of recommendation systems, who used to work for Yahoo and now works for Google

Representative literature: Matrix Factorization Techniques For Recommender Systems

Steffen Rendle

Rendle’s HomePage

Main contribution: The author of classic recommendation algorithm FM and BPR, now works for Google

BPR: Bayesian Personalized Ranking from Implicit Feedback

Hao Ma

Personal HomePage: HaoMa’s HomePage

Major contributions: a leader in the field of social recommendation, he has proposed many effective algorithms based on social recommendation

Microsoft literature: SoRec: Social Recommendation Using Probabilistic Matrix Factorization

Julian McAuley

Personal homepage: McAuley

His research interests include social networking, data mining, and recommendation systems. He is currently an assistant professor at the University of California, San Diego

Representative literature: Leveraging Social Connections to improve personalized ranking for Collaborative Filtering

Guo Guibing

Guibing Guo’s HomePage

Major contributions: a champion of recommendation system in China, founded LibRec, an open source recommendation system project

Representative literature: TrustSVD: Collaborative Filtering with Both the Explicit and Implicit Influence of User Trust and of Item Ratings

Hao Wang

Personal HomePage: HaoWang’s HomePage

Main contribution: Good at using deep learning technology to improve the performance of recommendation system

Representative literature: Collaborative Deep Learning for Recommender Systems

He Xiangna

Personal Homepage: Xiangnan He’s Homepage

Main contribution: Using deep learning technology to improve the performance of recommendation system

Representative literature: Neural Collaborative Filtering

Robin Burke

Personal HomePage: Rburke’s HomePage

Main contribution: Mixed recommendation direction of the bull

Representative literature: Hybrid Recommender Systems: Survey and Experiments

A bright

Main contribution: The expert in the field of domestic recommendation system with equal emphasis on theory and practice, won the second Prize of Netflix Prize

Representative literature: Recommendation System Practice.

Xie Xing

Xing’s Page

Major contributions: Focus on data mining, social computing, etc., good at interpretable recommendation research, etc.

Representative literature: A Survey on Knowledge Graph-based Recommender Systems

Jiliang Tang

Personal Homepage: Jiliang’s Page

Major contribution: Good at using social network analysis techniques to improve recommendation performance.

Social Recommendation: A Review

Zhao Xin

Zhaoxin’s HomePage

Main contribution: Famous scholar of recommendation system in China, focusing on improving top-N recommendation performance by using natural language processing technology

Representative literature: Improving Sequential Recommendation with Knowledge-enhanced Memory Networks

ishikawa

Personal HomePage: Shichuan’s HomePage

Main contribution: The research direction is recommendation on heterogeneous information network, proposed the similarity calculation of weighted heterogeneous information, etc

Representative literature: Semantic Path based Personalized Recommendation on Weighted Heterogeneous Information Networks

gogaku

Wu Le’s HomePage

Main contributions: The research direction is the combination of social information recommendation, proposed the neural influence diffusion model, etc

Representative literature: A Neural Influence Diffusion Model for Social Recommendation

Wang Hongwei

Personal Homepage: Hongwei’s Page

Major contributions: A focus on graph machine learning, with a focus on making recommendations in conjunction with knowledge graphs.

Representative literature: Multi-task Feature Learning for Knowledge Graph Enhanced Recommendation

Iii. Related papers

Review of collaborative filtration

Comparison of collaborative filtering algorithms: Limitations of current techniques and proposals for scalable, high-performance recommender systems

Mixed recommendation review

Hybrid Recommender Systems: Survey and Experiments

Social Recommendation Overview

Social recommendation: a review

Cross-domain recommendations are reviewed

Cross Domain Recommender Systems: A Systematic Literature Review Deep learning based Recommender system A survey and new perspectives

4. Related books

Recommendation System Practice

Recommender Systems Handbook

Recommender Systems :An Introduction

Recommendation Systems: Technology, Evaluation, and Efficient Algorithms

V. Related courses

Recommender Systems Specialization

Recommender Systems Specialization (Recommender Systems) was recently published on Coursea. This course will start on March 26, 2018. This course consists of four courses and one graduation project course, including introduction to recommendation system, nearest Neighbor collaborative filtering, recommendation system evaluation, matrix decomposition and advanced technology.

6. Recommend common data sets of the system

1.MovieLens

Applicable to traditional recommendation tasks, three different sizes of data are provided, including user rating information on movies, user demographics, and movie description characteristics.

2,Filmtrust

It is suitable for social recommendation tasks with a small scale and contains users’ rating information on movies as well as trust social information between users.

3,Douban

It is suitable for social recommendation tasks with a moderate scale, including users’ score information on movies and trust social information between users.

4,Epinions

It is suitable for socialized recommendation tasks with a large scale and contains users’ score information on movies as well as social trust information between users. It is worth noting that this data set also includes information of distrust relationship.

5,Yelp

It is almost applicable to all recommendation tasks, and the data scale is large, so you need to manually extract the information you need, including evaluation and scoring information, user information (registration information, number of fans, friends information), product information (attribute information, location information, image information), advice information, etc.

6,KB4Rec

This data set links the items in the recommendation data to the entities in the large-scale knowledge graph, providing the items in the recommendation system with structured knowledge information with rich semantics.

Vii. Code and tools

1.LibRec

Java version of the open source recommendation system, including more than 70 classic recommendation algorithms.

2,Surprise

The Python version of the open source recommendation system contains a variety of classic recommendation algorithms.

3,LibMF

C++ version of the open source recommendation system, the main implementation of matrix decomposition based recommendation algorithm.

4,Neural Collaborative Filtering

Python implements neural collaborative filtering recommendation algorithm.