An Improvement over Hybrid Recommender System Using Unsupervised-Supervised Learning

Abstract

Recommender systems are the functional behaviour and essential utility integrated within all online applications. These services control and suggest the product and services to users by acquiring their interest and usage behaviour. Various recommender systems were proposed and improved by the researchers for enhancing the accuracy of rank prediction. In this paper, the unsupervised and supervised learning methods are integrated within two layers of hybrid recommender system for optimizing the behavior of hybrid recommender system. The content-based grouping of items and users is optimized using Fuzzy-clustering. In collaborative filer, the Bayesian probabilistic analysis is performed for predicting the movie ranking. The comparative analysis is done against the conventional and improved collaborative and hybrid recommender systems. The results show that the proposed model reduced the RMSE and MAE significantly.