计算机集成制造系统 ›› 2023, Vol. 29 ›› Issue (8): 2801-2812.DOI: 10.13196/j.cims.2023.08.025

• • 上一篇    下一篇

基于离群点检测的关键顾客需求识别

曹雪静1,王宇1,曹进2,张娜1,李玉鹏1+,侯路遥1   

  1. 1.中国矿业大学矿业工程学院工业工程系
    2.常州星宇车灯股份有限公司
  • 出版日期:2023-08-31 发布日期:2023-09-12
  • 基金资助:
    国家自然科学基金资助项目(51505480,72001203);江苏省自然科学基金资助项目(KYCX21_2477)。

Identification of key customer requirements based on outlier detection

CAO Xuejing1,WANG Yu1,CAO Jin2,ZHANG Na1,LI Yupeng1+,HOU Luyao1   

  1. 1.Department of Industrial Engineering,School of Mines,China University of Mining and Technology
    2.Changzhou Xingyu Automotive Lighting Systems Co.,Ltd.
  • Online:2023-08-31 Published:2023-09-12
  • Supported by:
    Project supported by the National Natural Science Foundation,China (No.51505480,72001203),and the Natural Science Foundation of Jiangsu Province,China(No.KYCX_2477).

摘要: 对关键顾客需求进行分析和研究是获取产品(再)设计信息的有效途径。在考虑顾客需求重要度动态性的背景下,提出一种基于离群点检测的关键顾客需求识别方法。首先,基于邻域粗糙集理论确定顾客需求项的邻域关系;根据顾客需求项邻域关系所包含的知识确定顾客需求项之间的相似程度进而构建邻域信息网络。然后,结合马尔科夫随机游走策略构建关键顾客需求评估指标,并根据稳态向量识别关键顾客需求。最后,以某汽车产品的关键顾客需求识别为例,通过对比分析验证了方法的可行性和有效性。

关键词: 关键顾客需求, 离群点检测, 邻域粗糙集, 复杂网络

Abstract: Analysis and research on the key customer requirements is an effective means to obtain product (re)design information.Under the circumstance of considering the dynamic of customer requirements’ importance,an identification method of key customer requirements based on outlier detection was proposed.The neighborhood relationships of customer requirements were determined based on the neighborhood rough set theory.The similarity of influencing factors between customer requirements was determined according to the knowledge contained in the neighborhood relationships of customer requirements to construct the neighborhood information network.An evaluation index for the key customer requirements was constructed by combining the Markov random walk theory,and the key customer requirements were identified based on the stable vector.Taking the identification of key customer requirements of the automotive products as an example,the feasibility and effectiveness of the method were verified through comparison and analysis.

Key words: key customer requirement, outlier detection, neighborhood rough set, complex network

中图分类号: