Bootstrap Confidence Interval for Robust Cpk Index

A Cpk index is the most widely used measure of process capability. A better understanding and interpretation of Cpk is by constructing its confidence interval. The construction of such intervals, are normally based on the assumption that the measurements process can be treated as samples from a normal distribution. Nevertheless, many processes are not normal and have a heavy tail distribution, which probably due to the presence of outliers in the data. An alternative approach is to use a bootstrap confidence interval estimate of Cpk index which is based on robust method. The advantage of this estimate is that its usage does not depend on any distribution. A numerical example is presented to evaluate the performance of the bootstrap confidence interval of the robust Cpk index. The result indicates that the bootstrap confidence interval estimate based on the robust Cpk is better than the classical Cpk index.

Preview PDF artikel_5_vol1_no2.pdf Download (292kB)

Actions (login required)

View Item