Abstract: In order to obtain objective and reasonable results of students’ evaluation of teaching, this paper adopts an intelligent approach based on rough set theory. Rough set method can only process decision table containing given decision attribute. However, during the evaluation of teaching, the decision attribute is usually lost because of lacking objective scales in practical measurement. Pointing at this problem, decision attribute values are acquired from experts’ evaluation data set based on the method of Kruskal maximum tree fuzzy clustering, and the integrated decision table is designed by combining the students’ evaluation data set with the decision attribute data. Objective weight values of all evaluation indexes are obtained under the information entropy method based on rough set theory, and the evaluation is finished for the candidates. Example analysis and contrastive experiments with other existing evaluation methods are given to demonstrate the effectiveness and superiority of the proposed new method.

Key words: rough set theory, conditional information entropy, students’ evaluation of teaching, weight of evaluation index, Kruskal maximum tree fuzzy clustering

摘要： 为实现更为客观合理的学生评教，基于粗糙集方法进行智能化分析。粗糙集方法必然涉及到分析含有决策属性的决策表，而实际学生评教中由于缺乏客观的尺度评定教师的教学质量，造成相应决策属性的未知性。借鉴督导专家评价的优势，基于Kruskal最大树模糊聚类方法对专家评价数据予以划分来获取决策属性，与学生评教数据集组合，构造完整的决策表。基于粗糙集方法从信息熵的角度来客观求取各评教指标的权重值，完成对待评教教师的决策评价分析。实例分析及对比实验证明了方法的有效性和优越性。

关键词: 粗糙集, 条件信息熵, 学生评教, 评教指标权重, Kruskal最大树模糊聚类