Paper Title
Integrated Autism Detection using Behavioral Screening and Image Analysis Through Hybrid AI Models

Abstract
The timely identification of Autism Spectrum Disorder remains a significant challenge in clinical practice, with conventional diagnostic methods often lacking scalability and introducing subjective biases. We present a dual-module computational framework designed to address these limitations through the integration of behavioral screening and visual analysis. Our approach employs a primary analysis component that processes standardized behavioral metrics through several machine learning classifiers (Logistic Regression, XGBoost, and Support Vector Classifier), while a complementary experimental component explores visual pattern recognition through Convolutional Neural Networks. Testing was conducted on 1,055 participants with 17 behavioral indicators extracted from standardized screening tools. The behavioral analysis module demonstrated exceptional performance, with Logistic Regression achieving 100% accuracy on the test dataset. Though still in developmental stages, the visual analysis component showed encouraging results in extracting meaningful patterns from non-verbal cues. By combining these complementary analytical methods, our framework represents a step toward multi-modal ASD detection that could potentially improve screening efficiency while maintaining diagnostic reliability. This work contributes to the growing body of research on computational approaches to neuro developmental disorder detection and offers insights into how diverse data sources might be leveraged to enhance early intervention opportunities Keywords - Autism Detection, Behavioral Screening, Machine Learning, Neural Networks, Diagnostic Tools.