ANALISIS KAPABILITAS PROSES DENGAN FUNGSI DENSITAS KERNEL; PROCESS CAPABILITY ANALYSIS WITH KERNEL DENSITY FUNCTION

Setiyarti, Nuning, Danardono

2015 | Skripsi | FMIPA

Process capability analysis can not be separated from the statistical quality control analysis. Quality control analysis using standard methods usually requires the assumption that the quality characteristic of interest follows a normal distribution. Departures from this assumption can result in wrong control limits estimation. Nonparametric approach based on kernel density function is an alternative that can be used to perform quality control analysis when the assumption of normality is not met and has been proven to produce better control limits estimate. Furthermore, process capability analysis can be carried out with the same approach as well. The approach based on kernel density function is used to analyze elongation data of thread type 30 TR 1008 Cop and 30 RT SIRO KEPYUR Cop. The kernel function used is specified in the form of Epanechnikov kernel and bandwidth selection method used is unbiased cross-validation. Kernel control chart built based on kernel cumulative distribution estimate. Estimation of process capability is calculated by analyzing the process failure rate. From the analysis conducted with the specification limits determined by the company, it is known that the production processes of both thread types by the attention to quality characteristics elongation have very poor process capability. While the comparison of the process capability analysis results based on a kernel density function, assuming that the data follow normal distribution, as well as result of Box-Cox transformation of the elongation data of 30 TR 1008 Cop thread for several variations of specification limits can be concluded that the process capability analysis based on kernel density function performs better in analyzing the nonnormal quality characteristic for wider specification interval so that the quality characteristic observations are also a lot more within the interval.

Kata Kunci : Kernel density function, Unbiased cross-validation method; Kernel individual control chart; Process failure rate; Process capability