基于特征选择技术的顾客需求重要度确定方法 - 计算机集成制造系统.pdf

基于特征选择技术的顾客需求重要度确定方法  ∗ 耿秀丽，叶春明 (上海理工大学 管理学院，上海 200093)    摘要：针对专家决策和层次分析法（AHP ）等方法计算顾客需求重要度具有的主观性，根据不同配置的产品服 务组合方案需求属性及整体满意度数据，采用基于条件概率的特征选择技术识别关键顾客需求，计算其基本重要度， 并基于规则分析需求的 Kano 属性用于重要度调整。顾客需求竞争性重要度分析通常基于企业间的相对比较，忽视了 本企业顾客需求的实际表现，准确性差。结合最大偏差方法和需求的实际表现计算顾客需求的竞争性重要度，然后 依据顾客需求的 Kano 属性予以调整，最后采用加权算术平均方法整合顾客需求基本重要度和调整后的竞争性重要度 计算最终重要度。以挖掘机产品服务组合方案开发过程中顾客需求重要度的计算为例，验证了所提方法的有效性。  关键词：顾客需求重要度；特征选择；条件概率；竞争分析；最大偏差方法  中图分类号：TH122 ；N94         文献标识码：A  Determining the importance weights of customer requirements based on feature selection technique GENG Xiu-li, YE Chun-ming (Business School, University of Shanghai for Science of Technology, Shanghai 200093, China) Abstract ：To deal with the subjectivity in determining the importance weights of customer requirements (CRs) using experts decision, AHP, etc., a new approach based on feature selection technique using conditional probabilities was proposed to identify key CRs, determine the basic importance ratings of CRs and analyze CRs’ Kano attributes to modify importance ratings, which was based on data consisting of CRs’ characteristics and overall satisfaction degree of product-service concepts with different configurations. The current approaches for determining competitive priority ratings of CRs were based on comparison among companies, ignoring the real CRs’ performance of the researched company, and had lower accuracy. The maximal deviation approach and CRs’ real performance were combined to determine competitive priority ratings of CRs, and then the ratings were modified considering CRs’ Kano attributes.