In the industry, control charts are powerful statistical tools for process control. Shewhart control charts assume a normal distribution for the quality characteristic. Based on a sample of size n, if ‘tn’ is a statistic, Shewhart variable control charts have control limits of \(E(t_{n} ) \pm \,3\,S.E(t_{n} )\) . However, in the case of a skewed population, control limits must be constructed using an alternative method. To address this, several researchers have employed a skewness correction method to construct control charts for mean. This paper adopts the popular probability model, Odds Exponential Log Logistic Distribution (OELLD), to construct a control chart with skewness correction. Using the coefficients of skewness based on Quartile’s and Kelly’s method, we construct variable control charts for the mean of subgroups under OELLD. Coverage probabilities are also calculated using simulation techniques. The findings are then compared with those obtained from OELLD and other existing models.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Skewness Corrected Control Charts for Mean: OELLD

  • Pushpa Latha Mamidi,
  • Naga Durgamamba Arigela,
  • P. Mallikharjuna Rao

摘要

In the industry, control charts are powerful statistical tools for process control. Shewhart control charts assume a normal distribution for the quality characteristic. Based on a sample of size n, if ‘tn’ is a statistic, Shewhart variable control charts have control limits of \(E(t_{n} ) \pm \,3\,S.E(t_{n} )\) . However, in the case of a skewed population, control limits must be constructed using an alternative method. To address this, several researchers have employed a skewness correction method to construct control charts for mean. This paper adopts the popular probability model, Odds Exponential Log Logistic Distribution (OELLD), to construct a control chart with skewness correction. Using the coefficients of skewness based on Quartile’s and Kelly’s method, we construct variable control charts for the mean of subgroups under OELLD. Coverage probabilities are also calculated using simulation techniques. The findings are then compared with those obtained from OELLD and other existing models.