MULTI-AGENT E-LEARNING SYSTEM BASED ON ENHANCED UNIVARIATE AND PREDICTIVE EXTRA TREE TECHNIQUES FOR SUSTAINABLE DEVELOPMENT

Abstract

The multi-agent system is an effective system to inculcate knowledge through online mode. In this research work, two feature selection techniques, namely enhanced univariate and predictive extra tree have been proposed. These feature selection techniques are used to communicate between the multiple agents. The feature selection method proposed in this work is to predict the students' result. Machine learning algorithms have been employed to produce better results by selecting the relevant features from the database. The parameters evaluated are accuracy, precision, recall, and the F measure. The random forest algorithm has produced better results during the parameter analysis and the naïve bayes algorithm has produced comparatively poor results. Thus, the random forest is the optimized one for the proposed e learning multi agent system.