Machine Learning Models for Predicting Young Stroke Risk Using Clinical Data

Background Stroke in young adults, though less common than in older populations, carries major long‐term morbidity and socioeconomic impact. Predictive tools for early risk stratification in this group remain limited. Methods Clinical and biochemical data from young stroke patients were analysed, including vascular risk factors (hypertension, diabetes, alcohol use, autoimmune disorders) and laboratory parameters (hypocalcemia, hyperhomocysteinemia). After exclusion of incomplete records, data was split into training (70%) and testing (30%) sets. Seven supervised machine learning classifiers—logistic regression, decision tree, random forest, gradient boosting, support vector machine (SVM), k‐nearest neighbors (KNN), and naïve Bayes—were developed and optimized using five‐fold cross‐validation. Model performance was evaluated with accuracy, F1 score, and area under the receiver operating characteristic curve (AUC). Feature importance was assessed using model‐specific metrics and validated with permutation/dropout loss analysis. Results Among seven classifiers, KNN and naïve Bayes achieved the highest accuracy (88.9%) and F1 scores (0.888‐0.889). Boosting demonstrated excellent discrimination (AUC 0.915) despite moderate accuracy (83.3%). SVM (85.2%) and logistic regression (83.3%) also performed well, whereas decision tree yielded the lowest accuracy (77.8%). Across models, hypertension, alcohol use, and autoimmune disorders consistently emerged as the strongest predictors, with hypertension showing the highest dropout loss values (0.259‐0.314). Additional contributors included hypocalcaemia, diabetes, and hyper‐homocysteinemia. Conclusion Machine learning models reliably predicted young stroke risk using clinical and biochemical variables. KNN, naïve Bayes, and boosting offered the best predictive performance. Consistent identification of modifiable risk factors highlights opportunities for early intervention and personalized prevention strategies in young adults at risk of stroke.