Utilization of GC-MS in Determination of Electronic Nose Sensor Array for Classification of Gambung Green Tea Quality

Green tea is recognized for its numerous advantages. Several studies have been conducted employing diverse approaches to categorize the quality of green tea, one of which involves the development of an electronic nose system (e- nose). The selection of the gas sensor array is a critical factor in the construction of an effective e-nose system. Thus, in the present study, the identification of gas sensors array was executed based on the result of gas chromatography-mass spectrometry (GC-MS) on two specimens, categorized as good and quality defects. The sensory outputs from each sensor are gathered into a dataset, and subsequently analyzed for predictive classification. The dataset generated by the deployment of sensors, namely MQ3, TGS822, TGS2602, MQ5, MQ138, and TGS2620, exhibits optimal performance, specifically 0.989 in random forest modeling with regards to metrics such as accuracy, precision, recall, and F1 score. In summary, it is proven that employing a combination of these six sensors and random forest modeling yields a performance above 98%.