Chiang Mai Journal of Science

     Identifying meat species accurately is crucial for food safety, fraud prevention, and quality control in the food industry. Mislabeling or adulterating meat products can lead to economic losses and pose health risks to consumers. However, conventional methods for species identification, such as DNA analysis or spectroscopy, are often timeconsuming and expensive. In recent years, highly sensitive sensors have made estimating food product types and freshness possible. The BME688 sensor produced by Bosch Sensortec is one of the most sensitive gas measurement sensors today. In this study, various types of meat were classified using machine learning methods on the data obtained by the BME688 gas sensor. Each type of meat has a distinct microbiota composed of specific microorganisms that influence its spoilage process and the volatile compounds it releases over time. VOCs and VSCs released into the environment by microorganisms that develop over time in meat types can be detected with this sensor. In this study, Decision Tree, Gaussian Naive Bayes, Bagging Tree, Support Vector Machine, Xgboost, Logistic Regression, Multi-Layer Perceptron, And Bosch AI-Studio Neural Network models were trained and tested on data obtained from different meat types using BME688. The accuracy values of the trained models were compared, and it was determined that the GNB and BT models have the highest potential for possible usage scenarios. Considering the results obtained, it was revealed that the BME688 sensor could distinguish chicken, sheep, and cattle meat with near-perfect accuracy.