Recommender systems are software tools and techniques that provide users with recommendations of items. Nowadays, with the rapid growth of the web and the information on it, recommender systems have attracted much attention. Many reputable web sites use a recommender system for giving services to the users. Generally, recommender systems are composed of users and items in the way that each user has collected some items in its library. The recommender system, considering these items and the others that the users have collected, recommends the items that the users have not collected and may be interested in. By now many techniques have been used for developing the recommender. In this research we provide a new method for recommendation systems with collaborative filtering approach that have better results than studied methods. A particular dataset called MovieLens is created to assess the recommenders that the most recommenders are evaluated by this dataset. In this thesis we have compared the proposed method with Weighted Pearson algorithm which is user based and NBI, NBIW and INBI that have better evaluating result than other item based methods based on our current knowledge. The results according to the evaluation shows that the proposed method is better than these methods. Keywords: Recommendation Systems (RS), Collaborative Filtering (CF), Network-Based Inference (NBI), Bipartite Networks
