Classification of High-Achieving Students Using the Fuzzy K-Nearest Neighbors Algorithm Based on Academic History and Organizational Activities

Abstract

The selection of outstanding students in higher education is often conducted manually with a primary focus on academic performance, which may lead to subjectivity and inefficiency. This study aims to develop a classification model for outstanding students using the Fuzzy K-Nearest Neighbors (F-KNN) algorithm based on academic records and organizational activities. The dataset consists of 112 students from Universitas Muhammadiyah Makassar, including academic, achievement, and organizational activity data. The research methodology includes data preprocessing, feature engineering, feature weighting, application of the F-KNN algorithm, and model evaluation using confusion matrix and 5-fold cross validation. The experimental results show that the proposed F-KNN model achieved an accuracy of 91.3%, precision of 81.82%, recall of 100%, and F1-score of 90% using k = 5 and m = 2.0. These results indicate that F-KNN is effective and reliable for classifying outstanding students. This study contributes to the development of an objective decision support system for student achievement selection in higher education.