An Improved Personalized Collaborative Filterinng Algolrithm in E-Commerce Recommender System

Collaborative filtering recommender systems have become important tools of making personalized recommendations for products or services during a live interaction nowadays. However, there are still some drawbacks and challenges for CF based recommender system such as prediction accuracy, scalability and sparsity. This paper points out that from a certain angle, the predictions these systems produce are not really personalized ones which lead to the above problems. After the analysis of the traditional collaborative filtering algorithm, the authors then proposes a new personalized recommender algorithm based on traditional CF algorithm to improve the recommender system. At last the effectiveness and superiority of the proposed novel algorithm is proved by four experiments using both cosine correlation similarity and Pearson correlation similarity in this paper