Autism Classification Using Machine Learning Techniques

We are currently seeing a rapid spread of Autism Spectrum Disorder (ASD), so when studying autis0m behaviors, researchers note that this study requires substantial costs and time to characterize autism. Through the use of machine learning techniques, autism can be detected early. There are studies using machine learning techniques, but they have not provided any conclusion in determining the characteristics of autism due to the different ages of people. This study aims is to predict autism in any age group (children, adolescents, adults), using a classification system based on machine learning techniques (random forest (RF), decision tree (CART), Naive Bayes (NB), and Support Vector Machine (SVM). The results from algorithms (CART, SVM, NB, and RF) are evaluated using several metrics (Accuracy, Precision, Recall, F1 Score) based on the AQ dataset- 10. The model used showed advanced results in evaluating the accuracy of the types of datasets. The results provide superior performance for ASD classification. Random Forest and Support Vector Machine accuracy have been improved between (98% and 100 ^% ) with features selected by correlation technology and K fold split data.