A Low Cost Electronic Nose System for Classification of Gayo Arabica Coffee Roasting Levels Using Stepwise Linear Discriminant and K-Nearest Neighbor

Abstract

A low-cost electronic nose (E-Nose) system using metal oxide sensor (MOS) was developed for Gayo arabica coffee roasting level. The developed electronic nose was designed to have a simple, rapid detection, as wells as provides reliable results. The ENose system is equipped with MOS sensors, sensor chamber, microcontroller, computer, and data acquisition system. The level of coffee roasting was monitored by read the data from the sensors continuously in real time every second. The sensor signals were recorded in Excel file using data acquisition system and analysed using both stepwise linear discrimination and k-nearest neighbor classifiers. A high percentage (91.67%) of accuracy was obtained using stepwise linear discrimination method. Furthermore, k-nearest neighbor classifier using city block distance demonstrated higher accuracy than stepwise linear discrimination classifier. The results showed that the electronic nose system has a potential for assessing Gayo arabica coffee roasting level. The study confirmed that the proposed electronic nose equipped with at least two MOS sensors was suitable for monitoring the level of coffee roasting level. The result could be used for evaluating other varieties of roasted coffee.