Abstract—autism spectrum disorder (ASD) is a
developmental disorder that affects many people, especially
children, with problems in communication and social life.
Although many factors such as inherited gene mutation and
environment influences may play important roles in autism, the
actual causes of autism remain as myth. Without proper analysis
for ASD, many people cannot get early detection of ASD and the
public cannot fully understand ASD. Our goal is, through
scientific investigation, to increase the public awareness on
autism, so that better societal support could be provided to
individuals with ASD traits. In this research, we analyze a
children's autism screening dataset, apply feature selection
techniques to identify key characteristics associated with ASD
traits, and validate our findings with machine learning
predictions. Our research revealed social responsiveness factors
closely tied to ASD traits, and a strong link between autism
assessment questionnaires and ASD traits.
Index Terms—Autism spectrum disorder, categorical data,
feature selection, correlation, chi square test, mutual
information
Albert Zheng is with the Academy of the Canyons, Santa Clarita, CA,
USA.
He Zhu is with the Rancho Cucamonga High School, Rancho
Cucamonga, CA, USA.
Xinyi Hu is with the Los Osos High School, Rancho Cucamonga, CA,
USA.
Lan Yang, is with the California State Polytechnic Univ., Pomona, CA,
USA.
*Correspondence: anpei.zheng@gmail.com (A.Z.)
Cite: Albert Zheng, He Zhu, Xinyi Hu, and Lan Yang, "Using Feature Selection Techniques to Investigate the Myth of Autism Spectrum Disorder," International Journal of Future Computer and Communication, vol. 12, no. 4, pp. 93-99, 2023.
Copyright © 2023 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0)
Copyright © 2008-2024. International Journal of Future Computer and Communication. All rights reserved.
E-mail: ijfcc@ejournal.net