Principle Component Analysis (PCA) - Classification of Arabica green bean coffee of North Sumatera Using FT-NIRS

Abstract

Coffee is one of the important export commodities in Indonesia. Indonesia is among the top four coffee producers in the world after Brazil, Vietnam and Columbia. A total of 301 samples of North Sumatera Arabica green coffee bean were obtained from different districts namely Dairi, Humbang Hasudutan and Mandailing Natal. In this research, Principle Component Analysis (PCA) was used as to classify the three North Sumatra Arabica coffee bean, and also used several pretreatment smoothing data: derivative 1 (D1), derivative 2 (D2), Standard Normal Variate (SNV) and Multiplicative Scatter Correction (MSC) as spectra data correction methods. Result showed that the classification of the three North Sumatra Arabica coffee beans with PCA without pretreatment data was gaining 100% prediction accuracy.