Performance of A Portable NIR Spectrometer to Distinguish Coffee Species Based on Qualitative Chemometric and Artificial Neural Network (ANN) Models

Abstract

A wide range of genetic cultivars of coffee and their characteristics determine consumer preference and increase industrial actors’ awareness of production and marketing. The primary objective of this study is to develop a method to distinguish coffee species based on spectral characteristics acquired from a portable near-infrared spectrometer. The performance of this spectrometer in addressing classification problems is evaluated by the classification accuracy obtained from qualitative chemometrics, such as PCA and LDA, and artificial neural networks (ANNs) models. In this study, the instrument was successfully used and gained moderate accuracy for discriminating two coffee species, Arabica and Robusta, from Temanggung and Toraja. The accuracy was fair and achieved greater than 75%. Therefore, the instrument can be implemented as it provides simple, real-time, and in-situ analyses and can reach reliable results.