Using Adaptive Link Hiding to Provide Learners with Additional Learning Materials in a Web-Based System

Abstract

This article reports results from a web-based online learning experiment that provided learning support to students in an object-oriented programming course. This support was intended to assist learners in acquiring domain knowledge (in this case object-oriented programming knowledge) which they were to use later for problem solving. The course was delivered using adaptive support techniques in which the system interface adjusts in ways that suit different learners. The impact of using one of the implemented support techniques, adaptive link hiding, is reported here. Using this technique, the system provided links to additional learning materials according to its prediction whether or not a learner was likely to access them. The system’s decision was guided by a machine learning algorithm, the Naïve Bayes Classifier (NBC). The system’s prediction was compared to actual access of these additional learning materials, yielding a predictive accuracy of 72%.