K-Nearest Neighbors (K-NN) Algorithm Model in Predicting the Graduation Rate of Teacher Professional Education Students in Indonesia

Musthofa, Musthofa
Yunitasari, Dwi
Nasikhin, Nasikhin
Wang, Juanduo

International Journal of Social Learning (IJSL)
Vol. 4 No. 3 (2024): August

Predicting the graduation rate of the PPG program has an important significance in analyzing the factors that affect students' success in completing the PPG program. This study uses the K-Nearest Neighbor model in online learning to predict the pass rate of students in the Teacher Professional Education Program (PPG) at UIN Walisongo Semarang. The study analyzed data from 423 students, focusing on input quality variables, such as pedagogical competence and teaching innovation. Results showed the Wave 1 pass rate in 2023 was 86.7%, with 13.3% failure, a 1.7% decrease from Wave 3 in 2022. The confusion matrix showed significant improvement in True Positives (TP) and True Negatives (TN), with an accuracy of 0.916, precision of 0.3, and recall of 0.9725 students' academic achievement.