Modeling Students’ Growth in an Adaptive Learning System Using Deep-IRT

• In Event: Student Growth

Abstract

In this study, the Deep-IRT model is fitted to quiz items in an adaptive learning system. Preliminary results showed good model prediction accuracy. The next step is to integrate latent knowledge state estimates for the prediction of students’ growth on future assessments and to compare with traditional IRT methods’ outputs.

Authors

• Tongtong Zou

• Jinah Choi, EDMENTUM, INC.

• Sonya Powers, Edmentum, Inc.