• 论文 •    

基于预处理小波神经网络模型的协同创新客户评价与应用研究

杨洁, 杨育, 王伟立, 赵小华, 宋李俊   

  1. 1.重庆大学 机械工程学院,重庆400030;2.重庆通信学院 通信指挥系,重庆400035
  • 出版日期:2008-05-15 发布日期:2008-05-25

Evaluation of collaborative innovative customer based on PWNN model and its application

YANG Jie, , YANG Yu, WANG Weili, ZHAO Xiaohua, SONG Lijun   

  1. 1.School of Mechanical Engineering, Chongqing University, Chongqing 400030, China;2.Department of Communication Command, Chongqing Communication College, Chongqing 400035, China
  • Online:2008-05-15 Published:2008-05-25

摘要: 为了在协同产品创新中有效地识别和评价创新客户,提出了运用基于预处理的小波神经网络模型,对协同创新客户进行评价。在分析面向客户学习效应评价过程的基础上,建立了包括学习效应在内的,由五个方面构成的协同创新客户综合评价指标体系。运用粗糙集理论对评价指标进行预先处理,减少了冗余指标项,降低了小波网络的输入维数,采用迭代梯度下降法和逐步检验法确定小波网络结构,然后应用小波网络进行协同创新客户综合评价。应用结果表明了该评价模型的有效性和可行性。

关键词: 协同产品创新, 学习效应, 创新客户, 识别与评价, 小波网络

Abstract: To identify the innovative customers and evaluate their ability of innovation, based on Pretreatment Wavelet Neural Network (PWNN), a new model of collaborative innovation customer selection and evaluation was proposed. Firstly, the influence of customer learning potential was analyzed and the evaluation index system of collaborative innovation customer, which including customer's innovation knowledge, ability of innovation, collaborative attitude, learning potential, innovation requirement of customer, was constructed. Secondly, theory of rough set was utilized to simplify the customer evaluation index system and reduce the input dimensionality of Wavelet Neural Network (WNN). Then, algorithms of stepwise checkout and iterative grads descending were used to decide the parameters of WNN and to get the synthetic evaluation value of customers. Finally, a numerical example was used to illustrate the feasibility and effectiveness of this model.

Key words: collaborative production innovation, learning potential, innovation customer, identification and evaluation, wavelet neural network

中图分类号: