Comparison of Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) in the Old Classification of Case Settlement in Religious Courts

Abstract

The length of case settlement is one of the important aspects in the evaluation of the judicial process because it can provide an overview of the efficiency of case resolution. This study aims to compare the performance of the Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) algorithms in classifying the length of case settlement in the Religious Court into three categories, namely Fast, Normal, and Slow. The research data comes from the Religious Court case dataset consisting of 15,522 records and is then aggregated based on the case number so that data at the case level is obtained. The settlement time is calculated based on the difference between the registration date and the date of the verdict, then used as the basis for the formation of categories. The SVM and KNN models were tested using training and testing data with a ratio of 80:20, and hyperparameter tuning was carried out using 5-fold cross-validation. The evaluation was carried out using accuracy, precision, recall, and F1-Score. The test results showed that the KNN after tuning gave the best performance with an accuracy of 55.88%, precision of 55.40%, recall of 52.39%, and an F1-Score of 52.48%. The SVM after tuning obtained an accuracy of 53.35% and an F1-Score of 50.16%. These results showed that KNN Tuned was superior to SVM Tuned in the dataset used, although both models still faced difficulties in identifying the Slow category.