Combining regularization and logistic regression model to validate the Q-matrix for CDM

• In Event: Cognitive Diagnostic Models II

Abstract

In this study, we combine the L1 regularization and multiple logistic regression (MLR) model to validate the Q-matrix. Results show that the regularized MLR method: (a) produces better Q-matrix recovery rate, true positive rate, and true negative rate than either the MLR-B or the Hull method for most conditions.

Authors

• Xiaojian Sun

• Tao Xin, +8613552946058

• Naiqing Song, Southwest University