项目名称: 动态和多元非参数控制图的研究与应用
项目编号: No.11201246
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 数理科学和化学
项目作者: 李忠华
作者单位: 南开大学
项目金额: 22万元
中文摘要: 本项目将紧密结合国家《质量发展纲要(2011-2020年)》,对统计过程控制图的理论和构造进行深入研究,并期望为提高我国工业产品质量提供指导与帮助。本项目的主要研究内容有:(1)部分或者全部参数未知时基于p值控制图的构造;(2)基于动态抽样策略控制图的构造,考虑Box-Cox变换函数族;(3)非参数控制图的构造,考虑Liu (1995)的数据深度方法,Zou and Tsung (2011)的空间符号方法和Zhou et al (2009)的变点探查方法等;(4)高维数据控制图的构造,考虑Janic-Wròewska and Ledwina (2000)和 Ducharme and Ledwina (2003)的两样本检验方法,Feng et al. (2012)的稳健高维稀疏方差阵估计方法和Chen and Qin (2010)的超高维两样本检验方法等;(5)平均运行长度的计算。
中文关键词: 控制图;动态;多元;非参数;p值
英文摘要: The original and creative ideas of this project are represented as follows: (1)Control charts using the p-value approach when some of the process parameters or none of the parameters are known; (2) Control charts using dynamic sampling, i.e., finding much simpler and more effective function form from Box-Cox transformation family to obtain more effective sampling schemes than literature; (3)Control charts using nonparametric methods, such as the data depth approach of Liu (1995), the spatial sign approach of Zou and Tsung (2011) and the change point detection approach of Zhou et al. (2009), to construct schemes for monitoring mean shifts and variance shifts simultaneously or distribution-free schemes; (4)Control charts combining online detection methods with high dimensional data inference schemes, such as two sample test of Janic-Wròewska and Ledwina (2000) and Ducharme and Ledwina (2003), the robust estimation method for ultra high dimensional matrix of Feng et al. (2012), and two sample test for high dimensional data of Chen and Qin (2010); (5)Calculation of average run length.
英文关键词: control chart;dynamic;multivariate;nonparametric;p value