Abstract
The experience of drinking coffee is deeply impacted by the aroma of the beverage. Historically, human assessors have classified coffee aromas, but recent developments have enabled the use of electronic noses (E-noses) for a more consistent and automated approach. In this paper, we introduce a novel methodology that employs a Concentration Separability Indicator (CSI) to efficiently select features and enhance the classification process. By testing this method on a dataset of gases from two coffee brands at three different concentration levels, we have achieved significant improvements in classification accuracy, reaching up to 100% accuracy and a 100% F1 score. This breakthrough enhances sensor data analysis performance and has the potential to improve the efficiency and objectivity of coffee aroma classifica