Accurate Estimate of Autism Spectrum Disorder in Children Utilizing Several Machine Learning Techniques

Abstract

Autism Disorder is a neurologically proven disorder in which children have impaired communication and interaction abilities. Children with autism spectrum disorders are characterised by a lack of social engagement, repetitive behaviour, and unchanging interests. It is very important to diagnose it in the early stages of life. Nowadays, machine learning helps health care for such diagnoses, which also reduces the cost and time. The main goal of this research work is to imply different algorithms of machine learning in order to predict autism. In this study, the different classification methods for diagnosing ASD were used on children aged 4 to 11 years. The present study proposes the Support Vector Machine, K nearest neighbor, Decision Tree, and Linear Discriminant Analysis algorithms of machine learning to classify the autism spectrum disorder. A number of features are extracted from the data set using an algorithm and statistically analyzed. The dataset was separated into 70:30 ratios. A comparison of various performance measures was done after applying the mentioned algorithm. It is observed that decision trees and SVM give a higher accuracy of 99 percentage than KNN and LDA, with 70 percentage and 97 percentage, respectively. Python, which is commonly used for machine learning classifications, is used to calculate classifier performance.