On the learning curve attrition bias in additive factor modeling

Cyril Goutte, Guillaume Durand, Serge Leger

Abstract: Learning curves are a crucial tool to accurately measure learners skills and give meaningful feedback in intelligent tutoring systems. Here we discuss various ways of building learning curves from empirical data for the Additive Factor model (AFM) and highlight their limitations. We focus on the impact of student attrition, a.k.a. attrition bias. We propose a new way to build learning curves, by combining empirical observations and AFM predictions. We validate this proposition on simulated data, and test it on real datasets.