Autism Detection Using Artificial Neural Networks: A Comparative Study

Autism spectrum disorder (ASD) is a complex neurodevelopmental condition that manifests with varying degrees of social, communicative, and behavioral challenges. Early and accurate diagnosis of ASD is crucial for timely intervention and support. In this study, we explore the efficacy of Artificial Neural Networks (ANN) as a tool for automating and enhancing the process of autism detection. The objective of this research is to make a comprehensive analysis by comparison of ANN with other commonly used methods in the field of autism diagnosis. We leverage a well-curated dataset and implement a systematic approach to preprocess and classify autism-related features. Our ANN architecture, tailored to the unique characteristics of ASD, is rigorously evaluated alongside other machine learning and statistical methods. The performance metrics employed include F1-score, sensitivity, accuracy and specificity, allowing a thorough assessment of diagnostic accuracy. The results of this study reveal that ANN exhibits a notable advantage in the accuracy of autism detection when compared to alternative methods. Our findings not only underline the potential of ANN as a powerful tool for early diagnosis of autism but also shed light on the limitations and possibilities in the domain of automated ASD diagnosis. This research contributes to the ongoing efforts to refine the accuracy and efficiency of autism diagnosis, with practical implications for clinical practice, early intervention, and the well-being of individuals with ASD. The outcomes underscore the significance of considering ANN as a promising technology for enhancing the diagnostic process and call for further exploration of this approach in the field of neurodevelopmental disorders.