How can dense results be differentiated in comprehensive evaluations? A hybrid information filtering model

作者：

Highlights：

• An evaluator distance-based information filtering (IFED) model is proposed to differentiate dense evaluated results.

• The concept of feature distance is defined to quantitatively measure evaluation validity.

• The filtering method considers Spearman’s rank correlation coefficient and the proportion of valid evaluators as constraints.

• The IFED model enhances the discrimination and stability of the evaluation results in multiindex comprehensive evaluation.

摘要

•An evaluator distance-based information filtering (IFED) model is proposed to differentiate dense evaluated results.•The concept of feature distance is defined to quantitatively measure evaluation validity.•The filtering method considers Spearman’s rank correlation coefficient and the proportion of valid evaluators as constraints.•The IFED model enhances the discrimination and stability of the evaluation results in multiindex comprehensive evaluation.

论文关键词：Multiindex comprehensive evaluation,Information filtering,Student evaluation of teaching,Differentiation,Entropy

论文评审过程：Received 5 July 2021, Revised 25 October 2021, Accepted 27 October 2021, Available online 1 November 2021, Version of Record 13 November 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.107658