Machine learning demonstrates superior risk prediction for young stroke using clinical and biochemical factors, indicating potential for personalized prevention.
Key Points
KNN and naive Bayes achieved the highest accuracy of 88.9% in predicting stroke risk in young adults, highlighting the effectiveness of machine learning methods.
Among various predictors, hypertension and alcohol use were identified as critical factors in stroke risk prediction, emphasizing their clinical importance.
Analysis utilized seven machine learning classifiers, with boosting showing excellent discrimination (AUC 0.915) for young stroke risk prediction.
These findings suggest that early intervention based on identified risk factors could significantly improve outcomes for young individuals at risk of stroke.