基于模糊关联与BP网络的客户感性知识挖掘  

Applying fuzzy-association-rule techniques and artificial neural networks to customer Kansei knowledge mining

作　　者：石夫乾[1] 孙守迁[2] 徐江[2] 

机构地区：[1]温州医学院计算机系,浙江温州325000 [2]浙江大学现代工业设计研究所,浙江杭州310027

出　　处：《工程设计学报》2007年第5期349-353,358,共6页Chinese Journal of Engineering Design

基　　金：国家自然科学基金资助项目(60475025);教育部博士点基金资助项目(20050335096)

摘　　要：通过建立Web问卷调查系统获取用户对产品造型特征的感性反映信息,并对用户感性评价信息予以模糊表征,进行多维模糊关联法则挖掘,进而产生客户感性信息与产品造型特征关联规则高频项目集.利用BP神经网络的学习能力对不同时段关联规则进行训练、预测和整合,从而实现客户感性知识挖掘,为产品设计辅助与企划决策支持提供新思路.Customers＇ Kansei information on product geometry features is obtained through Web survey system. Customers＇ Kansei evaluation is in fuzzy expression and then multi-dimension fuzzy-association-rule technology is applied. And finally, high-frequency sets in which the geometry features of products are highly associated with customers＇ Kansei information are generated. Furthermore, the learning ability of BP neural network are utilized to train, predict and integrate association rules from different time. And then customers＇ Kansei knowledge mining is implemented, which provides new ideas for product aid design and enterprise decision-making.

关 键 词：模糊关联法则 倒传递类神经网络 产品设计 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]