Journal of Systems Engineering and Electronics ›› 2012, Vol. 34 ›› Issue (4): 827-832.doi: 10.3969/j.issn.1001-506X.2012.04.33

• 软件、算法与仿真 • 上一篇    下一篇

基于扩散先验分布的成组技术分类识别方法

唐宁, 蔡晋, 李原, 张开富   

  1. 西北工业大学现代设计与集成制造技术教育部重点实验室, 陕西 西安 710072
  • 出版日期:2012-04-25 发布日期:2010-01-03

Method of classify identification of group technology based on the diffuse prior distribution

TANG Ning, CAI Jin, LI Yuan, ZHANG Kai-fu   

  1. The Key Laboratory of Contemporary Design and Integrated Manufacturing Technology, Ministry of Education, Northwestern Polytechnical University, Xi’an 710072, China
  • Online:2012-04-25 Published:2010-01-03

摘要:

分类识别方法是实际应用中最为广泛的统计分析方法之一,特征识别是对对象几何模型进行解释,通过匹配几何特征部分与特征的形式描述来实现。提取产品特征信息是特征识别的难点。结合零件〖CD*2〗设备成组优化的网络模型,提出一种基于贝叶斯推理的扩散先验分布的识别算法。依据成组技术的零件分类编码系统对零件设备进行成组分类,通过扩散先验分布的贝叶斯推理分类识别方法,根据待判别样品的预报密度函数,建立后验概率比和分类识别规则,对待识别样本进行判别分类。

Abstract:

The classify identification is one of the most widely used methods of statistical analysis. The classify identification is used to explain the geometric model of object by matching geometric features and the description of features. To obtain the features of productions is the key point of the classify identification. In view of part equipment group optimized network model, a method of Bayesian inference is given. These studies classify the parts and equipments by the diffuse prior distribution of Bayesian inference classification. According to the judgment of forecast density function of sample, the posterior probability ratio and the identification rules to classify the samples are set.