Agarwood Classification using the Electronic Nose Method

Agarwood is an important commodity in the perfume industry, and its price depends on its aroma. The resin content is the main factor in the appearance of the distinctive aroma. The differences in agarwood aroma are difficult for human senses to distinguish. In this study, we developed an innovative agarwood aroma classification system using the electronic nose method. This system used a semiconductor gas sensor array to detect and measure the gas emitted by agarwood and a machine learning algorithm for automatic classification purposes. The results of the experiments show that the gas sensor array can provide specific patterns for each type of agarwood, namely Merauke, Bintuni, Sulawesi, and Borneo. The support vector machine (SVM) algorithm can differentiate each type of agarwood with an accuracy of 93.7%.