An Asymmetric Weighting Schema for Collaborative Filtering.

Several key applications like recommender systems need to determine the similarities between users or items. These similarities play an important role in many tasks, such as discovering users with common interests or items with common properties. Most of the traditional methods are symmetric which means that they always assign equal similarity to each user even when one user rating profile is completely like the other but not conversely. In this paper we combine the traditional methods with an asymmetric weighting schema to distinguish between the similarities of two users. Several experiments have been performed to compare the performance of proposed method with traditional similarity measures.