Survey on Hybrid Multigroup Co-clustering Using Collaborative Filtering Model

Abstract: Recommendation systems play a vital role in filtering and modifying the preferred information. Recommender systems are classified into three categories such as collaborative filtering, content-based filtering and hybrid filtering. Generally Collaborative filtering is a technique which typically utilized to construct personalized recommendations on the internet but it suffers from data sparsity and scalability problem. Few websites that uses the collaborative filtering technology include Amazon, Netflix, iTunes, IMDB. The majority of the clustering-based Collaborative Filtering model utilizes only historical rating information in the clustering procedure but disregard other data resources in recommender systems such as the social connections of users and the correlations between items.This paper presents a survey of existing techniques with the novelties highlighting the need of personalized recommendation techniques based on clustering and collaborative filtering. In proposed system we generate recommendations in an effective manner with comparatively better accuracy and least cost by using parallel and distributed approaches with the heterogeneous information sources to overcome the problems like data sparsity and scalability which are very common in recommender systems.

Keywords: Recommender systems, collaborative filtering, co-clustering, information fusion, data sparsity.