Stacked Machine Learning Framework Identifies Early Signs of Autism in Young Children with up to 99% Accuracy

Written by

in

Researchers developed a data-driven stacked ensemble machine learning framework that identified early signs of autism spectrum disorder in screening data for children aged 12 to 36 months, with accuracy reaching 99%. Based on the Q-CHAT-10 questionnaire, the framework uses SMOTE to balance the class distribution to 50:50, narrows the features from 18 to 10, and then integrates five base classifiers. It outperformed traditional single-model screening tools across several independent benchmark datasets.

Source:BIOENGINEER.ORG · Read Original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *