Perbandingan Antara T2 Hotelling dan Analisis Komponen Utama pada Pengendalian Proses Statistika (The Comparison of T2 Hotelling and Principal Component Analysis of Statistic Process Controll)

Oki Dwi Purwati (Jurusan Matematika, Universitas Sriwijaya)

Jurnal Penelitian Sains
No 15 (2004)

T2 Hotelling have been used succesfully as a multivariate statistical process control tool for detecting fault in processes with correlated variabels. In the present work, principal component analysis (PCA) is proposed for improvement of monitoring performance, i.e. PCA finds linier combination of  variabels that describe major trends in data set, furthermore PCA can be used ro reduced dimensionality. The number of principal component used for contruction is usually smaller than that original variabels. The result show that althought the reability of PCA is the same as T2 Hotelling methode in many cases or simulation data, but reduction variabels make the process be simple.