ANALYSIS AND DETECTION OF AUTISM SPECTRUM DISORDER BY USING MACHINE LEARNING TECHNIQUES

Autism Spectrum Disorder (ASD) is a developmental condition that affects communication and social interaction, typically emerging in early childhood. This study investigates the use of machine learning algorithms—including Naïve Bayes, Support Vector Machine, Logistic Regression, K-Nearest Neighbors, Neural Networks, and Convolutional Neural Networks (CNN)—to predict ASD in children, adolescents, and adults. Three publicly available ASD screening datasets were used, containing 292 (children), 704 (adults), and 104 (adolescents) instances respectively. After handling missing values and preprocessing, CNN outperformed all other models, achieving prediction accuracies of 99.53% for adults, 98.30% for children, and 96.88% for adolescents, indicating its high effectiveness.