Abstract
Autism, a neurodevelopmental disease is very crucial to detect early in childhood for a person to live a better life. In this present study an machine learning based approach for autism detection has been used on dataset and the comparative study has been done for better accuracy. In the dataset k-nearest neighbour (KNN), Support Vector Machine (SVM), Logistic Regression has been applied. Logistic Regression models have outperformed on the dataset than the other models. Next, for better understanding of the significant feature SHAP algorithm, an explainable AI method has been applied, SHAP demonstrates the impact of feature, which are affecting more for getting the accuracy.

This publication has 12 references indexed in Scilit: