A neurological condition known as an autism spectrum disorder (ASD) profoundly impacts an individual's lifelong ability to engage and interact with others. ASD is an illness that can be recognized at early stages in a person's life and is often classified as a "behavioral disease" due to the presence of numerous symptoms that frequently manifest within the first two years of life, as per what most people believe about autism theory, these challenges typically emerge during childhood and endure into adolescence and adulthood [1]. In response to the growing popularity of medical diagnosis, machine learning techniques [2] are used to help doctors make better decisions about a person's health, extensive efforts have been made to leverage various algorithms, including Naive Bayes, Support Vector Machine, Logistic Regression, K-Nearest Neighbors (KNN), Neural Networks, and Convolutional Neural Networks (CNNs) [3], to predict and analyze ASD-related issues across different age groups—children, teenagers, and adults. These endeavors have been detailed in a research survey paper. These predictive models were assessed using the ABIDE dataset, which is openly accessible for research purposes. The study's findings underscore the effectiveness of CNN-based prediction models, which consistently outperformed other machine-learning techniques [4]. Notably, these models achieved impressive accuracy rates of 99.53%, 98.30%, and 96.88% for screening and diagnosing Autistic Spectrum Disorder in datasets representing adults, children, and adolescents, respectively
Read full abstract