A Hybrid Ontology-based Recommender System Utilizing Data Enrichment and SVD Approaches

Abstract

A recommender system is a method of filtering data that provides a personalized recommendation list to a user where the user is interested. The semantic relationship from the ontology modelling does help to boost the accuracy of the recommender system based on recent research. In this paper, we propose a hybrid method to predict the unknown rating in the user-item matrix by using the semantic information of the ontology. The rating prediction utilizes the combination of user-based and item-based techniques. The predicted ratings boost the information of the input data of the model used in the recommender system as input data quality plays an important role in constructing the model. Experimental results demonstrated that the proposed approach achieves greater accuracy as compared to the baseline and existing methods.