A comprehensive approach to classifying Austim spectrum disorder: Evaluating feature selection strategies

Identifying individuals with Autism Spectrum Disorder (ASD) is challenging; therefore, effective feature selection and classification methods are essential. This study introduces an innovative feature selection strategy that employs the Wrapper method in conjunction with various classification algorithms, including Decision Tree (DT), Support Vector Machine (SVM), K-Nearest Neighbours (KNN), Logistic Regression (LR), and Naive Bayes (NB). Discriminative features are identified using hierarchical clustering, enhancing the efficacy of subsequent classification techniques. Benchmark ASD datasets are utilized to implement cross-validation processes that assess the effectiveness of the suggested methodology regarding accuracy, precision, recall, and F1-score metrics. by employing a graphical depiction alongside the previously specified performance metrics to evaluate the model.