Digitalization of Disability Athlete Management System Using Decision Tree Cart Algorithm to Identify Potential Athletes in NPCI Kediri

This research addressed inefficient data management and subjective assessment in disability athlete selection by developing an artificial intelligence-based system for the National Paralympic Committee of Indonesia (NPCI) Kediri. The study employed Research and Development methodology to create a digitalized platform that integrated a centralized data management system with a Decision Tree CART algorithm enhanced through standard deviation-based data augmentation. The research utilized a dataset of 72 disability athletes categorized into four types: Daksa Lower, Daksa Upper, Cerebral Palsy, and Tuna Daksa. Data collection encompassed anthropometric measurements (height, weight), sociodemographic information (age, gender), disability classification, and performance test results (40-meter sprint, 600-meter run, arm hang, sit up, vertical jump). The system implementation included data collection, augmentation to address limited sample size, algorithm implementation with weighted balance parameters for class imbalance, model validation, and web-based interface development. Rigorous testing employed 10-fold cross-validation across multiple data splits (90:10, 80:20, 70:30), with the 80:20 split achieving perfect consistency (100% accuracy, 0% standard deviation). Disability type emerged as the primary discriminative feature, creating homogeneous groupings with pure Gini values at terminal nodes. The implemented system transformed subjective talent identification into an evidence-based process while streamlining administrative workflows through responsive interfaces tailored to various stakeholder needs. Keywords: Decision Tree CART, Data Augmentation, Athlete Disability Management System, Artificial Intelligence