AUTISM DISEASE PREDICTION USING MACHINE LEARNING

Abstract

Autism Spectrum Disorder (ASD) affects millions of children globally, with early detection playing a vital role in ensuring timely intervention and improved outcomes. This study presents an integrated system for ASD prediction based on age-specific questionnaires, image-based behavioral analysis, and a severity assessment mechanism. The system begins with a user login and age categorization (under 3 years or 4–11 years), followed by questionnaire-based prediction. After the initial prediction, users are prompted to upload a child’s image for further analysis. If autism is detected, a severity evaluation is conducted using 10 targeted questions. Based on severity, the system recommends hospitals globally (70), hospitals within the user’s country (50), or home therapy (30). If no autism is detected, users receive a message indicating a clean report. This holistic, user-friendly approach supports parents and clinicians with AI- powered insights, enhancing early detection and personalized care pathways.