基于贝叶斯理论的质量控制图异常模式识别  

Control Chart Abnormal Pattern Recognition Based on Bayesian Theory

作　　者：侯世旺[1] 朱慧明[2] 

机构地区：[1]怀化学院商学院,湖南怀化418000 [2]湖南大学工商管理学院,长沙410082

出　　处：《统计与决策》2017年第10期18-21,共4页Statistics & Decision

基　　金：教育部人文社会科学研究青年基金资助项目(13YJC630049);中国博士后科学基金资助项目(2011M501272);山西省青年科技研究基金资助项目(2013021021-2)

摘　　要：在应用控制图实施质量控制的过程中,及时、准确地识别出控制图异常模式,对于引发过程异常的原因诊断以及消除意义重大。传统的识别方法需要预先获知模式的特征信息或大量的特定模式数据,制约了方法的应用。文章以绘制控制图的样本统计数据为基础,逐点计算各种模式参数的极大似然估计量,应用贝叶斯规则推算各模式出现的信度大小,并依据每个采样时点模式决策统计量的取值,对控制图模式做出判断。数值仿真实验验证了本文所提方案对于基本控制图模式在识别率与灵敏度方面的有效性。In the process of the quality control by using control chart, it is of great significance to recognize the abnormal patterns timely and accurately for the diagnosis and elimination of assignable causes. The traditional recognition methods require the pattern feature information or much of specific pattern data in advance, thus limiting the application. Based on the plotted data of control chart, this paper calculates point by point the maximum likelihood estimator of parameters for each pattern. At the same time the paper reckons the confidence degree of the occurrence of each pattern by using Bayesian rules. The control chart patterns are also estimated in the paper according to the decision statistic value at each sample point. The results of numerical simulations show the effectiveness of the proposed approach in accuracy and sensitivity for basic abnormal patterns recognition.

关 键 词：贝叶斯理论 控制图 异常模式识别 

分 类 号：O213.1[理学—概率论与数理统计]