A Machine Learning Framework for Early Detection of Autism Spectrum Disorder in Adults Using Feature Scaling and Classification Techniques

Autism Spectrum Disorder (ASD) is a neurodevelopmental condition that impacts communication, behavior, and daily functioning, where early diagnosis and support can greatly improve quality of life. However, detecting ASD in adults is challenging due to subtler symptoms and overlaps with other conditions. This study presents a machine learning framework to enhance adult ASD detection, employing feature scaling techniques such as Quantile Transformer, Normalizer, Power Transformer, and Max Abs Scaler to improve model performance. Eight classifiers—Random Forest, AdaBoost, K-Nearest Neighbors, Logistic Regression, Decision Tree, Gaussian Naïve Bayes, Support Vector Machine, and Linear Discriminant Analysis—were evaluated using metrics like Accuracy, Recall, Precision, Specificity, F1-score, and Area Under the Curve (AUC). Results showed that the Decision Tree model achieved 100% accuracy across all metrics on the adult ASD dataset. The findings underscore the importance of optimized machine learning techniques for accurate ASD detection, offering significant potential for early diagnosis and integration into healthcare systems.