Electronic Nose with Artificial Neural Network Method on Raspberry Pi 4 for Detecting Pork and Beef

As the country with the largest Muslim population, Indonesia promotes the adoption of a halal lifestyle, encompassing various aspects of life, including food. This research aims to differentiate between beef and pork using an E-Nose system, both in their pure form and when combined. The E-Nose system utilizes eight gas sensors (MQ 135, MQ 9, MQ 8, MQ 7, MQ 6, MQ 4, MQ 3, MQ 2) to measure the levels of gases commonly found in the process of distinguishing between beef and pork. The detected gases in beef and pork are converted into ADC signals by Arduino and then into PPM. The tested PPM values are stored in a database located on Raspberry Pi 4. Beef and pork are classified using the artificial neural network (ANN) method, and the analysis includes accuracy, precision, recall (sensitivity), true negative rate (specificity), F1-score, confusion matrix, and Principal Component Analysis (PCA). The predictions from the artificial neural network show correspondence with the training data. The results of this research indicate that the E-Nose system with the ANN method is able to differentiate between beef and pork with an accuracy of 99.77%. With this research, it is expected to effectively distinguish between beef and pork, both in their pure form and when combined, enabling individuals to avoid consuming forbidden haram food and addressing dishonest practices by traders.