聚类分析和因子分析在绩效考评中的应用

The Applications of Cluster Analysis and Factor Analysis in Performance Evaluations

摘要: 采用R软件编程，运用聚类分析和因子分析的方法来进行绩效考评。首先采用系统聚类分析法得到四种不同距离的谱系图及绩效考评分类表。其次采用因子分析选取教学因子和科研因子，利用回归法计算教师们的前两个因子得分，以及他们的两类综合得分。由第一类算法计算的综合得分平等地对待两个因子，由第二类算法计算的综合得分以因子的方差贡献率为权重进行加权。最后对教师绩效进行综合分析，得出对教师的绩效进行分类时，当无具体名额限制时，采用系统聚类分析法，当有具体名额限制时，采用因子分析法。按与系统聚类法得到的分类的相合性，得出因子分析的第二类算法优于第一类算法，并采用第二类算法的计算结果来对有名额限制时的教师进行分类。

Abstract: Programmed in the R software, we use the methods of cluster analysis and factor analysis for performance evaluations. First, we adopt the hierarchical clustering method to get the four different distance hierarchical graphs and the classification table of the performance evaluations. Second, we use factor analysis to choose the teaching factor and the researching factor, use the regression method to calculate the first two factor scores of the teachers, and their two kinds of composite scores. The composite score computed by the first kind of algorithm treats the two factors equally, and the composite score computed by the second kind of algorithm weights the two factors by their variance contribution rates. Finally, we do a comprehensive analysis of teachers’ performance evaluations and obtain that using the hierarchical clustering method when there are no specific quantitative restrictions, and using factor analysis method when there are specific quota restrictions. According to the consistency with the classification by the hierarchical clustering method, we obtain that the factor analysis of the second kind algorithm is better than that of the first kind of algorithm, and use the calculation result of the second algorithm to classify teachers with quota limitation.