Comparative Analysis of K-Nearest Neighbor and Support Vector Machine Methods for Assessing Quality Standards of Palm Oil Bunches

Abstract

Oil palm (Elaeis guineensis Jacq) is a crucial crop in the agricultural sector, particularly in Indonesia, as it produces various economically valuable products. The quality of oil palm fruit bunches (TBS) significantly influences the production process of crude palm oil (CPO), making accurate quality assessments essential for maintaining industry standards. This study aims to compare the effectiveness of two machine learning methods, K-Nearest Neighbor (K-NN) and Support Vector Machine (SVM), in determining the acceptable quality of TBS. Using TBS data from the years 2019 to 2023, the research analyzes several variables, including maturity level and yield percentage, to develop a web-based system for classifying TBS. The classification process involves preprocessing the data, applying the algorithms, and evaluating their performance based on key metrics such as accuracy, recall, and precision. The results indicate that the K-NN method outperforms SVM, achieving an accuracy of 100%, a recall of 100%, and a precision of 100%. In contrast, the SVM method demonstrates an accuracy of 91%, a recall of 100%, and a precision of 91%. These findings highlight the effectiveness of K-NN in classifying TBS quality while also demonstrating the reliability of SVM. This research is expected to provide valuable insights and effective solutions for decision-making regarding the acceptance of TBS quality, ultimately benefiting stakeholders in the palm oil industry and serving as a reference for future studies in data mining classification.