基于主成分分析构建指标权重的客观赋权法

Objective Weight Method Based on Principal Component Analysis to Establish Index Weight

针对主观赋权法在多指标评价过程中受人为因素的影响,导致评价结果随意性较强的问题,提出一种基于主成分分析构建指标权重的客观赋权法;将样本数据矩阵转置,依据客观数据之间的关系,用主成分分析的方法计算各项指标的得分,继而归一化处理,作为各项指标的权重系数。结果表明,该方法既能体现各项指标的相对重要性,又能体现指标间的差异性,将其应用于多指标综合评价中的评价结果能够达到科学、合理、客观的要求。

In view of the problem of strong randomness of multi-index evaluation results due to the subjective weight method influenced by human factors,an objective weight method was presented based on the principal component analysis to establish the index weight. The data matrix of samples was transposed,and the scores of each index were calculated by using the principal component analysis and the relationship among objective data. The scores were carried out in normalization processing to be regarded as the weight coefficient of each index. The results indicate that the weight coefficient can not only show the relative importance among evaluation indexes,but also show the difference. The method can be applied to the multi-index comprehensive evaluation,whose results can meet scientific,rational,and objective requirements.