GRAFIK PENGENDALI MULTIVARIAT ROBUST DENGAN ESTIMATOR MINIMUM COVARIANCE DETERMINANT UNTUK OBSERVASI INDIVIDU; ROBUST MULTIVARIATE CONTROL CHART USING MINIMUM COVARIANCE DETERMINANT ESTIMATOR FOR INDIVIDUAL OBSERVATION

Ike Fiandra Amrilian, Gunardi

2015 | Skripsi | FMIPA

To monitor a multivariate process , a Hotelling’s T2 control chart is often used. However it’s not rarely happens that a process contains outlying observations that can destroy control limit estimation and leads to masking and swamping effect. To solve this problem we use process control analysis using robust multivariate control chart using minimum covariance determinant estimator for individual observation. MCD estimator has high breakdown point and affine equivariant so it become highly robust estimator that can fight outlier observation without destroy limit control estimation. This thesis applies robust multivariate control chart using minimum covariance determinant estimator for individual observation to analyze diesel fuel production at fuel factory unit Pusdiklat Migas Cepu. By comparing robust multivariate control chart with standar multivariate control chart, it shows that robust multivariate control chart gives better conrol limit.

Kata Kunci : individual control chart; multivariate; robust; minimum covariance determinant; multivariate control chart; FAST-MCD.