Comparison of similarity measures to differentiate players' actions and decision-making profiles in serious games analytics. (November 2016)

• Record Type: Journal Article
• Authors: Loh, Christian Sebastian; Li, I-Hung; Sheng, Yanyan
• Is Part Of: Computers in human behavior. Volume 64(2016)
• Page Start: 562
• Page End: 574
• Publication Date: 2016-11
• Subjects: Serious games analytics -- Similarity measures -- Decision-making profiles -- Gameplay Action-Decision -- Partial Least Squares Discriminant Analysis (PLS-DA); Interactive computer systems -- Periodicals; Man-machine systems -- Periodicals; 004.019
• DOI: 10.1016/j.chb.2016.07.024
• Languages: English
• ISSNs: 0747-5632

Abstract

Abstract: Three Gameplay Action-Decision (GAD) profiles: Explorer, Fulfiller, and Quitter, have been identified based on individual's decision-making actions and navigational behaviors in situ serious games. The ability to profile trainees using serious games can yield new analytics and insights towards training and learning performance improvement, including the identification of weaknesses or potential training needs in the players towards adaptive training, and the creation of new diagnostics for prescriptive training, retraining, and remediation. Similarity measures of players' in-game course of actions (COAs) have been shown to be a viable approach in differentiating novices from experts in serious games. In this study, we examined and compared several popular similarity measures to see if any measure, or combination of measures, would be viable in differentiating players based on their GAD profiles in serious games. Our findings revealed that similarity measures, while significant in their predicting abilities individually, could gain more strength from one another in combination. More research is needed to create or develop new metrics and methods for players' action and behavioral profiling in Serious Games Analytics. Highlights: Similarity measures and gameplay action-behavior profiles are new research in Serious Games. Players can be profiled based on in-game navigational paths and decision-making strategies. Similarity measure is a viable method to differentiate player profiles by (dis)similarities. The use of n -gram is needed to preserve directionality of navigational paths in gameplay. Comparison of similarity measures will help identify new metrics and better analytics methods.