基于贝叶斯理论的质量控制图异常模式识别
发布时间：2019-04-08 18:30
【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。
[Abstract]:In the process of applying control chart to implement quality control, it is of great significance to identify the abnormal pattern of control chart in time and accurately, which is of great significance for cause diagnosis and elimination of process anomalies. The traditional recognition methods need to know the feature information of the pattern or a large amount of specific pattern data in advance, which restricts the application of the method. Based on the sample statistical data of drawing control chart, the maximum likelihood estimators of various model parameters are calculated point by point, and the reliability of each model is calculated by Bayesian rule. According to the value of each sampling point mode decision statistic, the control chart pattern is judged. Numerical simulations verify the effectiveness of the proposed scheme in terms of recognition rate and sensitivity for basic control chart patterns.
【作者单位】： 怀化学院商学院;湖南大学工商管理学院;
【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)
【分类号】：O213.1
[Abstract]:In the process of applying control chart to implement quality control, it is of great significance to identify the abnormal pattern of control chart in time and accurately, which is of great significance for cause diagnosis and elimination of process anomalies. The traditional recognition methods need to know the feature information of the pattern or a large amount of specific pattern data in advance, which restricts the application of the method. Based on the sample statistical data of drawing control chart, the maximum likelihood estimators of various model parameters are calculated point by point, and the reliability of each model is calculated by Bayesian rule. According to the value of each sampling point mode decision statistic, the control chart pattern is judged. Numerical simulations verify the effectiveness of the proposed scheme in terms of recognition rate and sensitivity for basic control chart patterns.
【作者单位】： 怀化学院商学院;湖南大学工商管理学院;
【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)
【分类号】：O213.1
【相似文献】
相关期刊论文 前10条
1 朱正森;;两类控制图不能混合一谈[J];轻型汽车技术;2002年08期
2 吴海英;合成控制图[J];武汉理工大学学报;2004年12期
3 王敏华;周娟;沈丹;;累积和控制图与休哈特控制图联合应用方法[J];湖北工业大学学报;2008年05期
4 王成斌;回归控制图[J];数理统计与管理;1988年01期
5 李跃波;合成控制图[J];数理统计与管理;2001年06期
6 庄常陵;控制图犯第Ⅰ/Ⅱ类错误分析[J];长春师范学院学报;2001年02期
7 张维铭;可变抽样区间的单边控制图[J];数理统计与管理;2002年06期
8 黄赢;控制图在检测实验室质量控制中的应用[J];现代测量与实验室管理;2002年04期
9 张公绪,孙静;统计过程控制与诊断 第四讲 ，
本文编号：2454812
本文链接：https://www.wllwen.com/kejilunwen/yysx/2454812.html
