Early Detection of Autism Spectrum Disorder Using Support Vector Machine and a Modified Grey Wolf Optimizer

Children with autism spectrum disorder (ASD) have difficulties with social interactions, communication, and sensory processing. For prompt support and intervention, early detection is crucial. This paper suggests a method for detecting ASD using Support Vector Machines (SVM) with hyperparameters optimized using a modified Grey Wolf Optimizer (mGWO), in light of the growing use of machine learning models in medical diagnosis. SVM performance is greatly impacted by hyperparameters like gamma and regularization (C). mGWO helps find the best values for these SVM hyperparameters, which enhances classification performance. By striking a equilibrium between exploration and exploitation, mGWO improves classification while overcoming the drawbacks of conventional GWO. The proposed method handles a variety of datasets containing clinical, genetic, and behavioral characteristics effectively. The proposed method gives better results in different open source datasets compared to other algorithms with specificity 1, precision 1, recall 1 and F1 score 1 in dataset 1 and specificity 0.9574, precision 0.94, recall 0.91 and F1 score 0.9247 in dataset 2 and specificity 0.9361, precision 0.89, recall 0.82 and F1 score 0.8536 in dataset 3.