Clinical Decision Support System for Autism Spectrum Disorder Screening Using Ensemble Machine Learning

Abstract

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder with a continuously rising global prevalence, where the WHO estimates that 1 in 100 children worldwide is affected by this condition. Early detection through screening instruments such as Q Chat-10 is crucial for appropriate intervention. This study aims to build and compare the performance of three machine learning algorithms Naive Bayes, XGBoost, and Random Forest in predicting ASD risk in toddlers. Model performance evaluation was conducted using a Confusion Matrix with accuracy, precision, and recall metrics. Experimental results show that the XGBoost algorithm delivered the best performance, achieving an accuracy of 98.10% and perfect precision (100%), making it the most reliable model for minimizing false positives. Meanwhile, Naive Bayes recorded an accuracy of 96.68% with superior sensitivity (recall 98.6%), and Random Forest yielded an accuracy of 95.73%. This study concludes that the integration of ensemble algorithms like XGBoost proves effective as an accurate clinical decision support method for early autism detection.