摘要
Abstract
Collaborative filtering recommendation algorithm is an important research direction in electronic commerce,the current collaborative filtering algorithms have lots problems such as the recommendation accuracy is low,algorithms are cold start.In order to improve the effect of collaborative filtering recommendation,a new algorithm is designed by combining user characteristic and item attributes.Firstly,the current study of collaborative filtering algorithms are analyzed to find out the causes of disadvantages,and then according to user characteristics and project properties,the similarity score is estimated,and according to the estimates the recommendation results can be obtained,finally the MovieLens data set is used to analyze the performance of collaborative filtering algorithm.The results show that the algorithm can solve the problems of the current collaborative filtering recommendation algorithms,and improve the accuracy of.It has better practical application value.关键词
电子商务系统/协同过滤推荐算法/用户特征/项目属性Key words
E-commerce system/Collaborative filtering recommendation algorithm/User characteristics/Project attributes分类
信息技术与安全科学
