Date of Award

12-2024

Document Type

Project

Degree Name

Master of Science in Computer Science

Department

School of Computer Science and Engineering

First Reader/Committee Chair

Jennifer Jin

Abstract

Autism Spectrum Disorder (ASD) diagnosis requires an integrative approach that combines behavioral, biomedical, and computational methodologies for enhanced accuracy. This study introduces a comprehensive framework that employs machine learning (ML) and deep learning (DL) techniques alongside linear regression to model relationships between behavioral traits, biomedical markers, and ASD likelihood. Behavioral inputs, such as social interaction patterns, repetitive behaviors, and communication characteristics, are analyzed using linear regression to identify significant predictors of ASD. Simultaneously, a Convolutional Neural Network (CNN) is trained on image datasets to detect visual cues, such as facial expressions, associated with ASD. Advanced techniques, including transfer learning and fine-tuning, are utilized to enhance the CNN's performance. By combining statistical and deep learning methods, this approach aims to deliver a robust and scalable solution for ASD diagnosis, promoting early detection and tailored interventions.

Share

COinS