IMPROVED WHALE OPTIMIZATION ALGORITHM BASED FEATURE SELECTION WITH FUZZY RULE BASE CLASSIFIER FOR AUTISM SPECTRUM DISORDER DIAGNOSIS

Autism spectrum disorder (ASD) is a neurological illness, which affects around 1% of the global population. A commonly clinical approach to diagnose ASD is the standardized ASD test that results in more diagnostic time period with raised medical expenses. An automatic ASD diagnosis models enable to identity and classify the ASD at the earlier rate. The rise of machine learning (ML) models has begun to show promising results on the diagnosis of neuropsychiatric illness. Keeping this in mind, this paper presents a new Improved Whale Optimization Algorithm (IWOA) based Feature Selection (FS) with Fuzzy Rule Base Classifier (FRBC) model, called IWOA-FRBC for ASD diagnosis. The presented IWOA-FRBC model employs IWOA based FS process to proficiently pick up the required set of features from the medical data. In addition, particle swarm optimization (PSO) algorithm is integrated to the classical WOA to attain better exploration capability. Besides, FRBC model is applied to determine the different class labels of ASD from the feature reduced subset. The effectiveness of the IWOA-FRBC model takes place by the use of 3 benchmark ASD dataset namely children, adolescent, and adult. The attained simulation results portrayed the goodness of the IWOA-FRBC model by attaining the maximum accuracy of 93.49%, 95.45%, and 94.23% on ASD-Children, ASD-Adolescent, and ASD-Adult dataset respectively.