Skripsi
PENCARIAN BAKAT (TALENT SCOUTINGI) RENANG PADA ANAK MENGGUNAKAN ALGORITMA K-NEAREST NEIGBORS
Swimming requires an objective and efficient talent identification system. This study develops a machine learning-based system using the K-Nearest Neighbors (KNN) algorithm to predict young athletes' potential based on anthropometric and motor skill data. The dataset consists of 50 normalized samples evaluated through K-Fold Cross Validation (K=5 and 10). Test results show the optimal configuration at K=2 with 60% accuracy for both validation methods. Although the overall accuracy remains below 65%, this system can serve as a preliminary tool for talent scouting, particularly for short and long distance categories. These findings indicate the need for further development in terms of data quantity and more optimal classification methods.