Weight Based KNN Recommender System

Today, the personalized recommendation is one of the most important technologies in the Internet and e-commerce system, along with the increasing number of users and commodities. Among personalized recommendation algorithms, CF (Collaborate Filtering) has been researched for many years. The similarity computation method, which is the key in personalized recommender, like cosine theorem or pearson correlation coefficient, does not consider the distinguish degree of the items. In this paper, we will propose weight Based similarity algorithm, called IR-IUF++, which updates pearson correlation coefficient. IR-IUF++ performs better than traditional similarity algorithm in our experiment.