Seafood Quality Detection Using Electronic Nose and Machine Learning Algorithms With Hyperparameter Optimization

Abstract

The fisheries sector holds great importance in the Indonesian economy, particularly in terms of its contribution to growth and development. Both the fisheries and marine industries play significant roles in driving economic activities. The rising consumption of marine fishery products has resulted in a growing market demand for high-quality and safe products. Meeting this demand necessitates the maintenance of freshness in marine fishery products. Thus, this research aims to develop a fast, cheap, accurate method utilizing an electronic nose (e-nose) and machine learning algorithms as an alternative method for assessing the freshness quality of marine fishery products (seafood). This experiment employs seven algorithms with hyperparameter optimization to obtain the best performance. Machine learning algorithms are used for classification and regression tasks. The objective is to detect the freshness quality of marine fishery products accurately (classification task) while also identifying microbial populations present in the seafood (regression task). Through extensive investigations, the classification and regression models, specifically employing the k-Nearest Neighbors algorithm, demonstrated remarkable performance, achieving a very high accuracy score. Furthermore, the regression model yielded an RMSE value of 0.03 and an R2 value of 0.995, indicating the effectiveness of the approach in assessing and quantifying the quality attributes of marine fishery products.