A Comparison of the Efficiency of Data Classification in Learning Factors through Open Educational System with Electronic Teaching Aids in Tertiary Level

Abstract

This research presented the result of a comparison of the efficiency of data classification in learning factors through open educational system with electronic teaching aids of tertiary level students. There were 3 primary techniques which were compared in this research as follows: 1) Random Forest technique, 2) Deep Learning technique, and 3) Naive Bayes technique. There were 10 attributes with 152,850 datasets of the data which were provided in electronic teaching aids of open educational system for tertiary students. The data was divided into 5 parts by Cross-validation Test method, called 5-fold cross-validation. The data was randomly sampled for each section which consisted of 30,570 dataset. There were 4 parts used to generate a model and the single part was utilized to examine the precision and accuracy of the model. The results revealed that the most efficient technique, used to classify learning factors through open educational system with electronic teaching aids of tertiary level students, was Deep Learning technique with 90.60 percent of precision, 98.39 percent of recall, 99.01percent of accuracy, and 0.321 F-measure at the acceptable level of evaluation.