Classification Of Students' Academic Competencies Using The K-Nearest Neighbor (Knn) Method At State Vocational High School 2 Medan

Abstract

The Vocational Competency Test (UKK) is an evaluation instrument to determine the level of competency mastery of Vocational High School (SMK) students. However, the process of determining student eligibility to participate in the UKK is still carried out conventionally, so it takes a relatively long time and risks causing inconsistencies in the assessment. This study applies the K-Nearest Neighbor (KNN) algorithm to group students' academic competencies into two classifications: Ready to Participate in the UKK and Not Ready to Participate in the UKK. The data used comes from a 15-item knowledge aspect questionnaire given to 78 students majoring in Electrical Power Installation Engineering (TITL) at SMK Negeri 2 Medan. The data is separated into 62 training data or 80% and 16 test data or 20%. The data processing stage begins with preprocessing through Min-Max transformation and normalization. Next, classification is carried out using KNN with K = 3 and Euclidean Distance as the distance measurement method. Model performance is analyzed through a Confusion Matrix based on Accuracy, Precision, Recall, and F1-Score. The test yielded an accuracy of 81.25%, precision of 76.92%, recall of 100%, and an F1-score of 86.96%. These findings demonstrate that the KNN algorithm can be effectively used to identify students' academic competency readiness levels for the UKK. Furthermore, the developed web-based system can assist teachers in identifying students' readiness levels more quickly, systematically, and objectively as a basis for decision-making in facing the Skills Competency Test (UKK).