Feasibility Study of Hybridization Schemes in Recommender Systems

Abstract

Nowadays, Internet is overwhelmed with vast amount of data which makes the process of selection of related information more and more difficult. One of the significant difficulties observed by a user is to discover relevant information as per their requirement or interest. Recommender system is an intelligent information filtering system which filters the overloaded information from past and present and gives the result based on the user’s personal interest. Principal techniques of RS are Collaborative Filtering (CF) and Content Based Filtering (CBF). CF is considered to be the best in all domains and always outperforms. Recent Research studies have focused on mixing recommendation methodologies including Collaborative, Content-based and Knowledge based methods in hybrid recommender system research in order to improve recommendation results. This shift in research focus calls for having a sound idea about basic recommendation algorithms and hybridization concepts, schemes and mixing sequence. The paper attempts to explore, analyze and interpret all possible combinations of hybridization methods that are viable both in design and implementation. Finally, precise remarks were able to be delivered upon redundancy and feasibility options of various hybrid recommender system combinations in burke’s recommender system space.