Implementation of Data Mining Using the K-Nearest Neighbor Method for Predicting Student Learning Achievement at SMPN 2 Kesamben

Improving the quality of education requires accurate predictive strategies
to identify student academic performance. This study aims to develop a predictive
model for student achievement using the K-Nearest Neighbor (K-NN) method with
data from SMPN 2 Kesamben. The model was constructed using academic
attributes (average report card scores), attendance, and extracurricular scores as
predictor variables. A total of 576 student records were divided into training and
testing datasets, analyzed using RapidMiner, and manually tested in Excel with
Euclidean distance and K=1.
The results indicate that the K-NN model achieved an accuracy of 96.96%,
with high precision in the “Good” and “Very Good” categories, although less
accurate in the “Poor” category. The inclusion of extracurricular scores
significantly enhanced prediction quality by providing additional insights into
student engagement. In conclusion, the K-NN method is effective for predicting
student performance and can support schools in making data-driven decisions for
more targeted academic interventions.