Classification of Palm Oil Quality Using Case-Based Reasoning Based on Odor and Optical Data

Palm oil is a widely used vegetable oil that has various applications in food, cosmetics, biofuels, and other industries. However, palm oil quality assessment is a challenging task that requires accurate and consistent methods to ensure its safety and suitability for different purposes. Current methods of quality assessment are often subjective, time-consuming, and costly, which limit their applicability and reliability. In this paper, we propose a novel approach to classify palm oil samples based on their odor and optical characteristics using case-based reasoning (CBR). CBR is a machine learning technique that uses past cases to solve new problems by finding the most similar cases and adapting their solutions. We collected data from 10 different quality levels of palm oil samples and extracted features related to their odor and optical properties. We then applied CBR to classify the samples based on their quality levels. The results were evaluated using confusion matrix showed that our CBR model achieved 97.23% classification accuracy, with two out of 10 samples being classified with 100% accuracy. This demonstrates the potential of CBR as a reliable, and cost-effective method for palm oil quality assessment. This research significantly contributes to improving quality control and assurance procedures in the palm oil sector through the development of an intelligent classification system that integrates e-nose and optical sensor data and could be applied to the autonomous process in this industry.