Data Mining the Factors of E-Learning Performance through Decision Trees and Apriori Associated Rules

Abstract

The purpose of this research is to find out the factors that influence E-Learning performance from the learners' perspectives. The research selected the data on students who were enrolled in a Life-Education course during one semester and did data mining from the web learning portfolios, learning styles and learning experiences using Apriori rules, Decision trees and Cross tables in order to perform statistical analyses. The web learning portfolios include the frequency of attending the class, pasting articles, on-line discussion, reading time, number of pages read, and reading progress. The factors of learning styles include active/reflective, sensing/intuitive, visual/verbal, and sequential/global. The learners' experiences include gender, types of software used, prior computer experience in years, average daily time spent on-line, and prior computer degree of capability. The dependent variables considered in this E-Learning evaluation are the logical and creative performance. We used Clementine software for data mining. In the Decision trees, this research found that the sensing style students could have better logical performance. From Apriori rules and Cross statistical analysis, the high average daily on-line time could create higher performance in creative measurements. Females had better performance in the creative parts of evaluation. Finally, interactive learning is the most important factor in E-Learning.