Autism Spectrum Disorder Detection Using Machine Learning Techniques

Abstract

A developmental disease called autism spectrum disorder (ASD) greatly reduces a patient's capacity for social interaction and communication in everyday situations. Using various machine learning strategies, necessitating (KNN)-K Nearest Neighbors, (LR)-Logistic Regression, (DT)-Decision Tree Classifier, (RF)-Random Forest Classifier, and (SVM)-Support Vector Machine. Using data taken from potential ASD patients’ medical records, a strong machine learning-driven strategy for autism detection is established. The ASD dataset, which is accessible to the public, is used to assess the suggested methods. There are 800 cases and 22 distinct attributes in the ASD screening dataset. The framework involves data collection, data visualization, data preprocessing, and implementation of machine learning model. A machine learning-based ASD detection system using logistic regression aims to determine if an individual has ASD or not in accordance with relevant features and behavioral patterns. Testing results are evaluated based on the performance metrics and the proposed system utilizes Logistic Regression which yields 0.85 accuracy, 0.78 precision, 0.76 recall, and 0.77 F1 score following comparison with the other models of machine learning. The suggested framework for ASD detection greatly streamlines and expedites the process of diagnosing ASDs. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.