基于ARMA模型的自相关过程能力分析与评价  

Analysis and Evaluation on Auto-correlated Process Capability Based on ARMA Model

作　　者：王斌会[1] 陈旭[1] 

机构地区：[1]暨南大学统计学系,广东广州510632

出　　处：《数理统计与管理》2011年第5期879-887,共9页Journal of Applied Statistics and Management

基　　金：教育部人文社科项目(07JA910004);广东省自然基金项目(8151063201000029)

摘　　要：传统的过程能力分析基于观测值彼此独立的假设,但是实际生活中的数据常呈现相关的状态。本文对一般的自相关模型进行过程能力分析,从理论和模拟两个角度研究了自相关对传统过程能力分析的影响。提出基于ARMA模型的过程能力分析方法,模拟和数值分析的结果表明对于相关过程而言,该方法较传统方法相比有很大的改进。The traditional analysis of process capability is based on the assumption that the observations are independent. But the actual data is often auto-correlated. This paper analysis process capability of the general self-related model, study the influence of auto-correlation on traditional analysis of process capability from theory and simulation. The method how to analysis process capability based on ARMA model is proposed. Results of simulation and numerical analysis show that the method we proposed has greatly improved to the traditional method if the data is auto-correlated.

关 键 词：相关 ARMA 过程能力分析 蒙特卡洛模拟 

分 类 号：TB114.2[理学—运筹学与控制论] O212[理学—概率论与数理统计]