Diagnosis of Autism using Machine Learning as a Healthcare Technology
Camellia Ray

Camellia Ray*, Computer Science, Birla Institute of Technology & Science Pilani, Hyderabad, India.

Manuscript received on May 03, 2020. | Revised Manuscript received on May 14, 2020. | Manuscript published on June 10, 2020. | PP: 290-298 | Volume-9 Issue-8, June 2020. | Retrieval Number: 100.1/ijitee.H6296069820 | DOI: 10.35940/ijitee.H6296.069820
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Autism is one of the inborn disease, researchers are presently focusing on. The autistic child faces inflexibility in language, thinking and behavior together with the difficulties in understanding emotional states of others. There are lot of interventions going on to make them understand the feelings of others and vice-versa. Now a day, ASD became one of the quick spreading diseases all over the world. Therefore there is a huge need to provide a time-consuming and easy accessible diagnostic tool to detect autism at an early stage to help the clinicians in providing prior medications. Though there is no proper curability of autism, still easy detection helps to provide better therapy session and supports the autistic child to lead a comfort independent life. The thesis deals with the building up of a model where the parents and relatives of a suspected autistic child can easily detect if they are suffering from autism by providing their answers of some particular questions related to the characteristics of autism. In order to build that model, the data were collected manually from different autism therapy centers in India and those raw data are then classified by using three different classifiers namely Logistic Regression, Support Vector Machine and Random Forest with Python as a programming tool to find out the one with higher accuracy by various analyses after pre-processing. The Random Forest classifier with the highest accuracy is utilized in framing the question based model for the early discovery of autism which can be operated as a primary diagnostic model to assist medical professionals technologically. 
Keywords: Autism, Diagnosis, Random Forest, Logistic Regression, Support Vector Machine, 10-fold cross validation.
Scope of the Article: Machine Learning