• 论文 •    

基于反向传播模型的知识生产绩效评估方法

赵新力, 任伟巍,   

  1. 1.中国科学技术交流中心,北京100045;2.中国科学技术信息研究所,北京100038;3.中国专利技术开发公司,北京100088
  • 出版日期:2007-07-15 发布日期:2007-07-25

Performance evaluation model of knowledge production based on BP neural network

ZHAO Xin-li, REN Wei-wei,   

  1. 1.China Science & Technology Exchange Center, Beijing100045, China; 2.Institute of Scientific & Technical Information of China, Beijing100038, China;3. China Patent Technology Development Company, Beijing100088, China
  • Online:2007-07-15 Published:2007-07-25

摘要: 为了揭示知识生产绩效评估问题,着重从订单型知识生产的特征入手,建立了订单型知识生产的绩效评估体系,确立了用于订单型知识生产绩效评估的反向传播模型结构。得到了各输入指标对绩效的影响程度及模型训练时权值变化的规律。最后,利用插值算法外延修正了权值,提高了模型精度,同时也对批量型知识生产绩效评估模型做了简要说明。

关键词: 订单型生产, 知识生产, 绩效评估, 反向传播模型

Abstract: Importance of evaluation to performance of knowledge production was discussed. Proceeding with the characteristics of MaketoOrder (MTO) knowledge production, the evaluation system was presented. The framework of Back Propagation (BP) neural network for the performance on MTO knowledge production was established. Contributive factors of indexes and the discipline of improvement of the weights in links were obtained. Precision of the BP neural network was revised by the method of interpolation extension, and the proposed evaluation model of batch knowledge production was also mentioned.

Key words: make-to-order production, knowledge production, performance evaluation, back propagation neural network

中图分类号: