Machine Learning Techniques Utilised for Autism Spectrum Disorder Detection

Abstract

A neurological developmental illness called autism spectrum disorder impacts a person’s ability to connect, communicate, and learn. Autism spectrum disorder influences individuals in distinct ways, as do the severity and its manifestations. Often, this illness is diagnosed in individuals between the ages of 1 and 5 years, and the symptoms include unique behaviours, interests, and potential social difficulties. In order tTo lessen the rise in instances of autism, it must be addressed at the beginning rather than later when it gets severe. Using various machine learning algorithms, autism may be predicted at a very early stage. In our proposed effort, we want to test and execute a variety of machine learning models as well as predict outcomes of autism diagnosed in children in the age range of 1 and 5 years. To analyse analyze the data from the previous decade, we followed a step-by-step procedure. The anticipated data of patients who have autism and those who do not be monitored as fresh data and utilised utilized to observe outcomes for upcoming patients. They leverage a range of methods involving machine learning like logarithmic regression (LR), Naive Bayes (NB), decision tree (DT), and k-nearest neighbors (k-NN) techniques to foresee autism spectrum disorder. After the former has been implemented, we expand our project to demonstrate some cutting-edge features in order to retain accuracy.