Autism Spectrum Disorder (ASD) is a neurological condition characterized by difficulties in communication and social interaction, with a growing global concern about its early diagnosis. Early detection plays a crucial role in improving the quality of life through timely intervention. While machine learning (ML) techniques have shown promise in ASD diagnosis, their practical application in clinical settings remains limited due to the lack of interpretability and explainability. This study aims to bridge this gap by exploring various machine learning models for ASD detection, including Stacking Classifier, Voting Classifier, XGBoost Classifier, and DecisionTree Classifier. Moreover, the study emphasizes the integration of explainable AI techniques to enhance the transparency of the models, thereby ensuring that clinicians can understand and trust the predictions. The comparative evaluation of these models highlights their efficacy in terms of both performance and explainability, offering a clinically interpretable approach for early ASD detection.

Keywords: Autism Spectrum Disorder, Early Detection, Machine Learning, Explainable AI, Stacking Classifier, Voting Classifier, XGBoost Classifier, DecisionTree Classifier, Model Interpretability, Clinical Application.