Enhancing Autism Screening Classification Using Feature Selection and Stacking Classifier

The current screening process for early detection of autism spectrum disorders (ASD) is time-consuming and costly., requiring numerous questions about various aspects of a child's development. To address this issue., this study integrates the Recursive Feature Elimination (RFE) method into a stacking ensemble classifier., allowing to identify the most important and effective features from the autism screening tool. This approach is aimed to create a simplified version of the autism screening and to make the screening process faster and more efficient by reducing the number of questions in autism screening tool. The proposed model provides a more efficient and simplified alternative for autism screening., allowing for early decision-making based on consistent and precise results. With 0.9760% accuracy results in predicting ASD traits., it is hoped that these findings will be an alternative option to make the screening questions much simpler while also providing an alternative to parents in predicting autism earlier and with less time.