SOFTWARE SYSTEM FOR RECOMMENDING STUDY MODULE ON COURSE COMPUTER AND CONTROL ENGINEERING

Abstract

This paper contains an performance analysis on two types of recommender systems. First type is a recommender system based on collaborative filtering and the other one is based on latent factor model. Semantic web technologies are used for storing data, data are stored on Fuseki server and they are accessed via SPARQL queries. Recommender system based on collaborative filtering have the accuracy of 83.92% while the accuracy of LFM recommender system is 30.04%. It was concluded that even though collaborative filtering is more primitive method because it uses the part of whole student grades dataset.