Int J Performability Eng ›› 2021, Vol. 17 ›› Issue (6): 504-510.doi: 10.23940/ijpe.21.06.p2.504510

Previous Articles     Next Articles

An Integrated System for Initial Prediction of Autism Spectrum Disorder

Azhagiri.M, Shubhanjay Mishra*, Shubham Joshi, and Amritash Srivastava   

  1. Computer Science & Engineering, SRM Institute of Science and Technology (Ramapuram Campus), Chennai, 600089, India
  • Contact: * E-mail address: shubhanjaymishra@gmail.com

Abstract: Autism Spectrum Disorder (ASD) neuro-developmental disarray incorporates numerous conduct issues like socializing and lack of communication. The traits of autism tend to show during early childhood and regularly last all through an individual's life. The symptoms include avoiding eye contact, needing to be separated from everyone else, and experiencing difficulty identifying with others or not having an interest in others. It also includes conditions like recursive examples of conduct, territories of interest, or exercises. A framework to record, identify and label the different patterns of conduct of people with ASD has been created. In this study, various sorts of classification methods for ASD diagnosis were used on the dataset consisting of people from different age groups. The classification techniques that we used were LDA (Linear Discriminant Analysis), Logistic regression and KNN (KNearest Neighbor). Using these methods, the findings suggest that the model performed best using Logistic Regression to predict Autism Spectrum Disorder with a precision of 99% during the validation trial.

Key words: Autism disorder, ADI-R, KNN, logistic regression, LDA, classification, AQ-Chat