Recommender System Using Clustering Based On Collaborative Filtering Approach

Recommendation process plays an important role in many applications as W.W.W. Recommender systems uses the users, items, and ratings information to predict how other users will like a particular item. An important response to the information overload problem is provided by the recommender system, as it presents users more personalized and practical information services. In the recommender systems field Collaborative Filtering (CF) is one of the most successful technique. For users based on their neighbor’s preferences Collaborative filtering creates better suggestions. Current CF suffers from poor accuracy, scalability, data sparsity and big-error prediction. As the number of users and items in recent years are growing rapidly poses some key challenges for recommender systems. In this paper, we borrow ideas of object typicality from cognitive psychology and propose a novel Object typicality-based collaborative filtering recommendation method (OTCO), It outperforms many CF recommendation methods on recommendation accuracy (in terms of MAE) in MovieLens data set, In our propose method main approach is to cluster the all items into several item groups by applying k-means clustering algorithm. To help users to search items more easily and to improve the accuracy and quality of the recommendation. KeywordsRecommender System, Collaborative Filtering (CF), Object Typicality, user group and item group.