Tracking knowledge proficiency of students with calibrated Q-matrix

作者：

Highlights：

• A new calibrated Q-matrix-based framework is designed for knowledge tracing.

• The relationships between exercises and knowledge points are deeply explored.

• Three Q-matrix calibration methods are proposed to eliminate the bias in Q-matrix.

• The diagnosis results are applied on three educational tasks.

摘要

•A new calibrated Q-matrix-based framework is designed for knowledge tracing.•The relationships between exercises and knowledge points are deeply explored.•Three Q-matrix calibration methods are proposed to eliminate the bias in Q-matrix.•The diagnosis results are applied on three educational tasks.

论文关键词：Knowledge tracing,Calibrated Q-matrix,High-order connectivity,Graph convolution network

论文评审过程：Received 25 March 2021, Revised 22 December 2021, Accepted 23 December 2021, Available online 30 December 2021, Version of Record 5 January 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.116454