Abstract
Memory-based collaborative recommender system (CRS) computes the similarity between users based on their declared ratings. The most popular similarity measure for memory-based CRS is the Pearson correlation coefficient which measures how much the two users are correlated. How ever, not all ratings are of the same importance to the user. The set of ratings each user weights highly: differs from user to user according to his mood and taste. This will he reflected in the user's rating scale. Accordingly, many efforts have been done to introduce weights to Pearson correlation coefficient. In this paper we propose a fuzzy weighting to the Pearson correlation coefficient which takes into account the different rating scales of different users so that the rating deviation from the user's mean rating is fuzzified not the rating itself. The experimental results show that Pearson correlation coefficient with fuzzy weighting outperforms the traditional approaches.