Electronic Nose Based on Sensor Array for Classification of Beef and Rat Meat Using Backpropagation Artificial Neural Network Method

Abstract

The differentiation of beef and rat meat is crucial for food safety and consumer protection. This research aims to create a tool to distinguish between beef and rat meat and to analyze the training data patterns for both types of meat. A sensor array consisting of three gas sensors—TGS822, TGS2602, and TGS2610—was used to detect the presence of Metal Oxide Semiconductor (MOS) gases in the meat samples. The classification method employed was a backpropagation artificial neural network (ANN). Results indicate that the classification tool performs well in differentiating beef from rat meat, with distinct patterns observed in the training data for each type of meat. The model achieved a precision of 100%, a recall (sensitivity) of 80%, and an accuracy of 90%. However, the TGS2610 sensor did not show a significant difference between beef and rat meat, suggesting no variance in the gas content detected by this sensor. These findings highlight the potential of using such sensors in practical applications for meat detection and underscore the need for further refinement in sensor selection and system integration to improve classification performance.