Improving the expressiveness of black-box models for predicting student performance
作者：
Highlights：
• Black-box classifiers are proposed to predict student performance.
• Black-box classifiers are powerful and generalizable but difficult to interpret.
• Some tips about their design are proposed to improve their expressiveness.
• Some graphical tools are proposed to exploit the expressiveness and help students.
摘要
•Black-box classifiers are proposed to predict student performance.•Black-box classifiers are powerful and generalizable but difficult to interpret.•Some tips about their design are proposed to improve their expressiveness.•Some graphical tools are proposed to exploit the expressiveness and help students.
论文关键词：Black-box models,Prediction,Student performance,Graphical representation
论文评审过程：Received 16 March 2016, Revised 27 July 2016, Accepted 1 September 2016, Available online 9 September 2016, Version of Record 26 April 2017.
论文官网地址：https://doi.org/10.1016/j.chb.2016.09.001
