Classification of Salt Quality Based on the Content of Several Elements in the Salt Using Machine Learning

Abstract

Salt is one of the commodities in Indonesia. Salt has a very strategic and sustainable role for human life. Apart from being used for daily consumption, salt is also used as a raw material for various industries Indonesia, as a country surrounded by coastlines, can be self-sufficient in salt production and meet domestic salt needs. However, not all the salt produced maintains sufficient quality for consumption. Therefore, monitoring of the produced salt's quality is necessary to categorize it. Even though the categorization of salt quality is still carried out manually, this research employs data mining techniques with three different algorithms: Naï ve Bayes, K-Nearest Neighbor (K-NN), and Support Vector Machine (SVM), to simplify and enhance the efficiency of the classification process. The dataset used was obtained from salt data in the Sumenep region of Madura that consists of 349 records with seven attributes: sulfate, magnesium, water content, calcium, not dissolved, NaCl(wb), and NaCl(db) with four data classes that represent grades of salt quality (K1, K2, K3, and K4), and the salt data is divided into training and testing sets using the k-fold cross-validation method. Test results indicate that the K-NN method provides better outcomes compared to other methods, with an AUC value reaching 99.0%, accuracy of 91.7%, F1 Score reaching 91.6%, precision of around 91.9%, and recall of around 91.7%.