Abstract
Autism spectrum disorder (ASD) is characterized by a wide spectrum of symptoms, spanning challenges in social interaction and repetitive behaviors. Globally, researchers are dedicated to unraveling the prevalence and underlying mechanisms of ASD. Biomarkers, serving as indicators of biological processes, play a crucial role in this endeavor. They ofer objective insights into the physiological irregularities of ASD, surpassing traditional diagnostic methods and facilitating treatment monitoring and personalized interventions. In our retrospective case-control study, which included 51 children diagnosed with ASD and 40 typically developing controls, we employed ensemble learning to rigorously evaluate the significance of biomarkers associated with ASD. We utilized various classifiers within the ensemble learning framework to enhance the robustness of our analysis.