Recognition of Original Arabica Civet Coffee based on Odor using Electronic Nose and Machine Learning
Whilly Harsono, Riyanarto Sarno, Shoffi Izza Sabilla
Abstract
Many studies have used an electronic nose (E-nose) to detect several types of coffee. To the best of our knowledge, none of the studies have tried to detect odors from a mixture of several types of coffee. Therefore, this research proposes E-nose which can be used to recognize original Arabica civet coffee. The mixture of Arabica civet coffee and Robusta coffee (non-civet coffee) is used as the object of this research. Nine combinations of mixture are prepared in this study. Those combinations are referred to as classes. After collecting the data, a statistical calculation would be determined to obtain parameter statistics. Moreover, the classification method used in this study is to recognize original Arabica civet coffee and original Robusta coffee. Several classifications had been compared, namely Logistic Regression (LR), Linear Discriminant Analysis (LDA), and K-Nearest Neighbors (KNN). The best result is the KNN method with an accuracy value of 97.7% for nine classes.