Prediksi Ketuntasan Siswa Berbasis Data Multidimensi Menggunakan Metode K-Nearest Neighbor (KNN) di SMK NU Hasyim Asy'ari 2 Kudus

Huda, Muhammad Syafi’ul
Mulyo, Harminto
Wibowo, Gentur Wahyu Nyipto

Jurnal JTIK (Jurnal Teknologi Informasi dan Komunikasi)
Vol 10 No 4 (2026): OCTOBER 2026

This research implements the K-Nearest Neighbors (KNN) algorithm to predict student learning mastery at SMK NU Hasyim Asy’ari 2 Kudus for the 2025/2026 academic year using multidimensional data. Following data preprocessing and labeling via median thresholding, the results indicate that the best performance is achieved at $K$ values of 7, 9, and 10, with an accuracy of 58.62%. While the precision of 0.69 demonstrates reasonable accuracy in predicting students who achieve mastery, the recall of 0.50 highlights the model's limitations in identifying all students who actually pass. These results are primarily influenced by the limited sample size and imbalanced class distribution. Overall, KNN serves as an effective initial approach for objective academic prediction, though further optimization through parameter tuning or feature engineering is required to enhance future accuracy.