Classification of kimchi using laser-induced breakdown spectroscopy and k-nearest neighbors modeling

Abstract

This study introduces a novel application of Laser-Induced Breakdown Spectroscopy (LIBS) combined with k-nearest neighbors (KNN) modeling to classify the origin of kimchi. Using the spectral intensities of Mg II at 279 nm and K I at 766 nm, we achieved a classification accuracy of 92.8 %. This method effectively leverages regional differences in salt supply chains impacting kimchi's elemental composition. The innovation lies in applying the interclass distance method for variable selection in LIBS analysis, enhancing the interpretability and accuracy of food classification. Compared to traditional elemental analysis techniques, LIBS offers a practical, cost-effective solution for rapid field analysis with minimal sample preparation. This study not only demonstrates the potential of LIBS for food authenticity but also provides insights for developing accurate methods for detecting Mg and K in various food products, contributing to advancements in food quality control.