采用主成分分析-信息熵法评价人机系统方案

Scheme Evaluation of Man-machine System Based on PCA & IE

为解决人机系统方案评价中指标众多、权重客观性难于保证等问题,提出一种基于主成分分析与信息熵的方案评价新方法。采用主成分分析法约简评价指标体系,对原始数据矩阵进行标准化以去除量纲,建立标准化数据矩阵的相关矩阵并求其特征值,利用特征值计算方差贡献率和累计方差贡献率并确定主成分的个数。同时,引入信息熵方法对约简后的指标体系进行客观赋权,通过计算评价指标的熵值和差异系数求得指标客观权重,进而求取各方案的综合得分并完成设计方案的评价。最后,以煤矿用防突钻机的方案评价为例,对提出的方法进行了验证。

In order to solve the problems such as mass indices and unguaranteed weights objectiveness, a new evaluating method based on PCA （Frincipal Component Analysis ） and IE （Information Entropy ） was presented. Evaluation index system was reduced by PCA, so the original data matrix was standardized and the dimension of indices was removed. By constructing the correlation coefficient matrix of standard data matrix, the eigenvalue were gained, and both variance contribution and cumulative variance contribution were calculated and also the number of principal components was determined. At the same time, IE was used to determine weights objectively. Through the calaulation of entropy values and difference coefficient, the objective weights of indices were gained further. And then, scores of each scheme could be calculated and evaluation process of schemes was finished. In the end, this method was validated by example of anti-outburst drilling machine＇s scheme evaluation.