Autism is a neurological disorder that has long-term consequences for a person's interactions and communiqué with others. Autism is a "behavioral disease" which can be identified at any age. Symptoms usually arise within the first 2 years of life. Ac-cording to the ASD issue, it begins in childhood and progresses through puberty and maturity. With the growing usage of ML approaches in medical diagnosis research, this work investigates the feasibility of employing SVM, LR, KNN, Decision Tree, Random Forest, and XGBoost for predicting and assessing ASD difficulties in adults. On a publicly available non-clinically ASD dataset, the offered approaches are tested. The ASD screening in adolescent dataset has 801 cases and 20 attributes. Following the application of several machine learning approaches and the treatment of missing values, the findings strongly imply that Logistic Regression-based prediction models perform better on this datasets with an accuracy of 88%.

Autism Spectrum Disorder (ASD), K-Nearest Neighbor (KNN), Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), XGBoost (XGB).