Open AccessPersonalized Recommendation via Suppressing by Users and Items
Author(s) -
Can Wang,
Kun Wang,
Tao Wei
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1237/4/042020
Subject(s) - movielens , novelty , popularity , recommender system , information overload , computer science , disadvantage , diffusion , collaborative filtering , range (aeronautics) , binary number , diversity (politics) , information retrieval , artificial intelligence , world wide web , mathematics , psychology , social psychology , physics , materials science , arithmetic , sociology , anthropology , composite material , thermodynamics
An efficient recommendation system is fundamental to solve the problem of information overload in modern society. In physical dynamics, material diffusion based on binary networks has a wide range of applications in recommendation systems. However, material diffusion has the disadvantage of excessive diffusion and excessive attention to high-prevalence items. Most of the previous studies focused on reducing the popularity of the item. This paper suppresses the excessive diffusion between people and objects by simultaneously adjusting popular users and items. It evaluates the algorithm through two real datasets (Movielens and Netflix), which proves the method is superior to other algorithms in accuracy, diversity and novelty.