Improving prediction accuracy of multi-criteria recommender systems using adaptive genetic algorithms
Mohammed Hassan, Mohamed Hamada · 2017 Intelligent Systems Conference (IntelliSys) · 2017
Recommender systems are powerful intelligent systems that are considered as solutions to the problems of information overload. They provide personalized lists of recommended items to users using some machine learning techniques. Traditionally, the existing recommender systems used single rating techniques to estimate users' opinions on items. However, as preferences of the users might depend on several items' attributes, the efficiency of the traditional single rating recommender systems are considered to be limited since it can not account for the various items' attributes. A multi-criteria recommendation is a new technique that uses ratings to various items' attributes to make more efficient predictions. Nevertheless, despite the proven accuracy improvements of multi-criteria recommendation technique, research needs to be done continuously to establish an efficient way of modelling the criteria ratings. Therefore, this paper proposed to use an adaptive genetic algorithm to model multi-criteria recommendation problems using an aggregation function approach. The empirical results presented in this paper have shown that the multi-criteria recommendation technique using adaptive genetic algorithm has by far provided more accurate predictions than traditional recommendation approach.