期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2015
卷号：30
期号：4
页码：174-181
DOI：10.14445/22312803/IJCTT-V30P131
出版社：Seventh Sense Research Group
摘要：Recommender systems (RS) serve as valuable information filtering tools for web online users to deal with huge amount of information available on the Internet. RS can be used in making decision in various fields like which books to purchase or which music to listen and so on. In this paper we have proposed and implemented an algorithm based on the Collaborative filtering method and Matrix Factorization technique SVD. Collaborative filtering is one of the traditional method for Recommendation Systems based on the user feedback. Matrix factorization is a method to address the problem of Sparsity. In this paper , first sessions are formed based on the timestamps of user logs. Collaborative filtering is used to form clusters based on users and items. SVD is applied for the useritem matrix formed from the clusters to address the Sparsity problem. Finally recommendations are given to the new test users by using user and item clusters. Experiments are performed on the benchmark data set for the proposed algorithm and results shows improvement of the recommendation system accuracy over traditional collaborative filtering method.
