Comparison of k-NN and Naive Bayes Algorithms for Classifying Mackerel Tuna Freshness Through Gas Sensors

The high production and consumption of mackerel tuna in Indonesia and globally underscore its significance both as a staple food and as an export product that contributes to the national economy. Mackerel tuna is valued for its nutritional content and affordability, making its freshness and quality critical for consumer satisfaction and trade. This study demonstrates the effectiveness of k-Nearest Neighbors (k-NN) and Naüve Bayes algorithms in classifying fish freshness by analyzing gases emitted during spoilage, using MQ-2, MQ-9, and MQ-135 gas sensors. Both models achieved accuracy rates close to 100%. The k-NN algorithm performed with near-perfect accuracy, misclassifying only one sample out of 7,207. Although the Naüve Bayes algorithm was slightly less precise, it maintained high accuracy while offering prediction times nearly seven times faster and requiring 400 times less memory. These findings provide valuable insights for the development of real-time quality control solutions using machine learning and gas sensor technologies.