Abstract:

Improving student productivity in online learning depends on designing learning environments based on principles derived from learning science research into how people learn.  Students master a skill by solving the sequence of practice exercises related to the skill.  The initial development of a skills model on a course, defining skills and associate them with exercises, heavily relies on human intuition, thus it might have a discrepancy which leads to differences between expected and actual student performance.

Learning Factors Analysis (LFA) is an approach for improvement of a skills model by introducing difficulty factors which split a skill.  Although LFA is a powerful tool, difficulty factors need to be defined by domain experts.  This talk will start with a short introduction to LFA and evaluating method of skills models.  Then we will talk about our ongoing research for automatic difficulty factors mining.