Robust Artificial Bee Colony Optimization Based Classifier for Prediction of Autism Spectrum Disorder

Abstract

Autism Spectrum Disorder (ASD) is a significant type of neurological disorder that affects the ability of person's to connect socially and communicate with others. Repetitive and limited patterns of conduct are also a part of ASD. The expenses of autism can spike while a diagnosis is being sought and when therapies are being delivered, but many of these expenditures are ongoing and will stay with a person for the rest of their lives. In order to get better outcomes, machine learning and optimization have extended across a wide variety of professions and specialties. In this paper, an optimization-based classification algorithm namely Robust Artificial Bee Colony Optimization based Classifier (RABCOC) is proposed for precise detection of ASD. RABCOC performs optimization before classification is done. For performing classification, this research work makes use of enhanced decision tree based gradient boosting method. RABCOC is evaluated with three different ASD screening dataset with the metrics Accuracy and F - Measure. Results achieved by RABCOC are compared with existing classifiers and it is found that RABCOC has better performance than existing classifiers towards prediction of ASD.