Used cooking oil is cooking oil that has been used repeatedly, causing its quality to deteriorate and potentially harm human health. Many people continue to use used cooking oil due to cost-saving considerations and the difficulty of directly assessing its quality. Therefore, a system capable of determining the suitability of used cooking oil is needed. This research aims to classify the suitability of used cooking oil based on clarity, viscosity, and color parameters using the Naïve Bayes method. The classification system utilizes an LDR sensor to detect clarity levels, a YF-S401 water flow sensor to measure viscosity, and a TCS3200 color sensor to read RGB values, with a NodeMCU ESP32 microcontroller as the processing unit. The test results show that the LDR sensor successfully detects clarity levels, the YF-S401 sensor achieves an accuracy of 97.73%, and the TCS3200 color sensor reads RGB values with accuracies of 99.85% for red, 97% for green, and 83% for blue. The dataset used in this study consists of 120 training samples and 30 test samples. The classification process using the Naïve Bayes method produced an accuracy of 96.67%, a precision of 94.74%, and a recall of 100%.
Copyrights © 2026