Rock sugar precise classification method by combining electrochemical property analysis and machine learning model

Despite advances in production, current rock sugar grading methods fail to capture subtle quality differences across product types. In this study, a precise method for rock sugar classification was developed by integrating machine learning with electrochemical characterization. Classification was based on correlations between physical attributes (color, opacity, smoothness, and sweetness) and electrochemical signals. Data were collected using a three-electrode setup with a copper film working electrode, platinum counter electrode, and saturated calomel electrode (SCE) reference. Electrochemical techniques employed were electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), differential pulse voltammetry (DPV), and chronoamperometry (i-t). A boosted support vector machine (SVM) ensemble achieved 96.0 % classification accuracy across 10 rock sugar types, while K-means clustering grouped them into 3 categories. Combining physical and electrochemical parameters significantly enhanced accuracy and reliability. The proposed method offers a robust solution for quality assessment, supporting improved control and automated grading.