The Design of a Rule Base for an e-Learning Recommendation System Base on Multiple Intelligences

Abstract

The problem of e-learning systems, learners are given learning contents that do not match individual aptitudes. This paper aims to design a rule base for recommendations focusing on e-learning and learning profiles which are based on multiple intelligences. Design of the rule base was divided into four sections as follows. The first section covered a survey of the variables. Second section was creation of the questionnaire. Third section was a survey of the student sample groups. The last section was an analysis of data generated from the results of the survey. The process of selection for the rule base was undertaken by comparing the performance of the following algorithms 1) ID3 algorithm 2) C4.5 algorithm 3) NBTree algorithm 4) Naï ve Bayes algorithm 5) Bayes Net algorithm. The C4.5 algorithm had the highest percentage of prediction. Percentage of prediction from the C4.5 algorithm equaled 83.436%.