Autism Prediction using Machine Learning

Autism Spectrum Disorder (ASD) is a lifelong developmental condition that affects how individuals interact, communicate, and behave. Detecting it early can make a big difference by allowing timely support and intervention. In this study, we explored how machine learning techniques specifically XGBoost, Random Forest, and Decision Tree can be used to predict ASD using data that includes medical history, behavior patterns, and demographic details. Before building the models, we cleaned and prepared the data by handling missing values, converting categories into numbers, and using a technique called SMOTE to balance the dataset. Among the models tested, Random Forest performed the best, with an accuracy of 93%. It also highlighted that factors like age, a family history of autism, and behavioral responses played a big role in prediction. Our findings suggest that machine learning can be a powerful tool in supporting early diagnosis of autism, making the process faster and more accessible in real-world healthcare settings.