Kajian Pengaruh Noise dalam Analisis Komponen Utama untuk Peubah-Peubah yang Berkorelasi
Nugrahanto, Fajrianza Adi
MetadataShow full item record
Principal Component Analysis (PCA) is one of multivariate techniques that generally used for dimension reduction. PCA uses covariance matrices as initial information. Change in values of those matrices can result in different PCA scores. One of conditions that can cause the change is noise presence, which was studied by Tsakiri and Zurbenko (2011). It showed that PCA results will be different when the data were affected by noise. This study was conducted to see the effect of noise on PCA results for data with correlated variables. The results showed that noise have greater influence on PCA results for certain correlation coefficient values.