Classification of Autism Using Feature Extraction Speed Up Robust Feature (SURF) with Boosting Algorithm

Autism is a complex disorder that affects an individual's neurobehavioral development. This disorder is characterized by restricted and repetitive communication patterns, difficulties in social interaction, and changes in sensory processing. This study aims to classify autistic and normal faces by using feature extraction using the SURF (Speeded Up Robust Features) method and the boosting algorithm. This study uses several variants of boosting algorithms, such as Adaboost, Gradient Boosting, and LightGBM. The dataset used in this study consisted of 100 samples of autistic faces obtained from Special Schools in Medan, as well as 100 samples of normal faces. Comparison of training data and data testing is 70%:30%. The results of this study indicate that the boosting algorithm provides good performance in classifying autistic and normal faces. Among the boosting algorithm variants used, Gradient Boosting achieved the highest accuracy of 91.67%, followed by LightGBM with 88.33% accuracy, and Adaboost with 81.67% accuracy. This finding demonstrates the effectiveness of the boosting algorithm in this classification task.