Determining the Distinctiveness of Learners with Outlier Detection Ensembles

Abstract

Gamification is the application of game design principles and elements to non-game contexts. It has been increasingly used to engage and motivate learners. A clear understanding of learners’ characteristics is a key success of gamification. Understanding the distinctiveness of the learners' behaviors enables gamification design to encompass all types of participants. This paper then proposes a methodology to assist gamification design in identifying distinctive individual learners within a group. The proposed methodology adopts ensembles of outlier detection techniques to examine how much individual learners differ quantitatively from the group, and utilizes decision tree classifiers to identify the factors contributing to the distinctiveness. The outcomes of these methods are presented in the if-then rules, which assist the interpretability of the discovered insights. This method enables educators and gamification designers to personalize gamified learning environments by focusing on unique learner characteristics.