A Comparative Analysis of Machine Learning Models for Prediction of Autism Spectrum Disorder Using Screening Data

Abstract

Autism spectrum disorder (ASD) is a neurological and developmental disorder that affects how people interact with others, communicate, learn, and behave. ASD prediction is difficult because the diagnostic factors may not be based solely on observation. In this research paper, an in-depth comparative analysis of various machine learning models applied to the task of classifying autism traits was presented. Our study aimed to assess the performance of these models within the context of identifying individuals with autism based on relevant features and data. The machine learning models investigated in this study encompassed Logistic Regression (LR), Random Forest (RF), Support Vector Machine (SVM), k-Nearest Neighbours (KNN), Naive Bayes (NB), and Neural Network (NN). The models were evaluated using six essential classification metrics: accuracy, precision, recall, specificity, F1 score, and AUC score. On the training dataset, our results reveal nuanced performance characteristics. SVM and RF excel in precision and recall, showing promise for accurate autism trait classification. KNN exhibits remarkable specificity, suggesting its potential for minimizing false positives. LR and NB demonstrate balanced performance across multiple metrics, while NN exhibits high precision and recall, albeit with higher computational demands. It was concluded that SVM was the best classification model for autism trait classification.