Analytics of Reading Patterns Based on Eye Movements in an e-Learning System
Wednesday, March 20 5:45 PM-7:00 PM Location: Celebrity Ballroom 1&2
Presider: Monique Woodard, Drexel University, United States
Abstract: We investigate relationships between a subject’s impression of difficulty and the reading behaviors of teaching materials based on eye movement data of students learning in an e-learning system. The analyses are important to help revising difficult contents in teaching materials. In this study, we use complex teaching materials containing text, pictures, mathematical expressions, while a previous study in the content difficulty focused on simple teaching materials. We define reading pattern codes to analyze temporal and spatial information of eye movement data using non-negative matrix factorization and compute a backward reading pattern score to compute correlations of the relationships. Our summarization could extract unique reading patterns to understand overall reading behaviors, and we observed correlations between a subject’s impression of difficulty and reading behaviors.