Abstract

Many recommender systems recommend items to users with using single-criteria recommend systems. In addition to this, multi-criteria recommend systems have become popular recently. New techniques, that is based on multi-criteria recommend systems is more accurate recommends than single-criteria recommend systems are developed. In this paper, with new techniques, we achieved more accurate results for the simple bayesian classifier for multi-criteria recommend systems. The simple Bayesian classifier is a very succesful supervised machine-learning algorithm. In our approach, we calculated similarity between users for each criteria with using binary similarity measures. After, we made classification with using the simple Bayesian classifier with various techniques. Our emprical results show that the simple Bayesian classifier based on multi-criteria rating system significantly improve classification accuracy.