An Intelligent Machine Learning and Deep Learning Framework for Early Prediction of Autism Spectrum Disorder

Abstract

Autism Spectrum Disorder (ASD) is a neurodevelopmental condition that can affect communication, social interaction, behavior, and learning abilities. Early identification of ASD is important because timely intervention can improve developmental outcomes and support personalized care. However, conventional diagnosis often depends on behavioral assessments, clinical observations, and specialist expertise, which can be time-consuming and difficult to access. This study proposes an intelligent framework for the early prediction of Autism Spectrum Disorder using Machine Learning (ML) and Deep Learning (DL) techniques. The proposed system processes relevant clinical, behavioral, demographic, and assessment-based features to identify patterns associated with ASD. Various machine learning algorithms are applied for classification, while deep learning models are explored to automatically learn complex relationships within the input data. Data preprocessing techniques such as missing-value handling, normalization, feature selection, and data balancing are incorporated to improve prediction performance. The models are evaluated using performance measures including accuracy, precision, recall, F1-score, and ROC-AUC. The proposed framework aims to provide an efficient and reliable computer-assisted approach for early ASD risk prediction, potentially supporting healthcare professionals in making timely decisions. By combining ML and DL methods, the framework seeks to improve predictive capability while reducing the limitations associated with traditional screening approaches.