Study on collaborative filtering recommendation algorithm based on web user clustering

摘要

Collaborative filtering recommend is the most widely used and the most successful recommendation algorithm. However, because the online effective amount of information on the number and types of goods is growing rapidly, to recommend system proposed a serious challenge, the collaborative filtering recommend exists in the cold start and sparse matrix, realtime problems need to be solved urgently. In order to solve the problem, this paper based on the collaborative filtering algorithm proposed Web recommend system based on user clustering, analysis of the Web recommendation system implementation process, and finally, experiment design and analysis. The results show that the proposed collaborative filtering recommendation based on user clustering method and the traditional collaborative filtering method is compared, and can efficiently improve recommendation quality, and better meet the needs of users.