An Ensemble Machine Learning and Data Mining Approach to Enhance Stroke Prediction.

Area of Science:

• Computational medicine
• Health informatics
• Machine learning in healthcare

Background:

• Stroke is a major global health concern, necessitating accurate early prediction for effective intervention.
• Current prediction methods require enhancement for improved accuracy and reliability.

Purpose of the Study:

• To develop and evaluate an ensemble machine learning approach for superior stroke prediction.
• To compare the performance of various machine learning models in identifying stroke risk factors.

Main Methods:

• Applied the CRISP-DM methodology with techniques like Random Forest, ExtraTrees, XGBoost, ANN, and GANN.
• Utilized SMOTE for dataset balancing and hyperparameter tuning via grid/randomized search cross-validation.
• Evaluated models using accuracy, precision, recall, F1-score, and AUC metrics.

Main Results:

• The ensemble ExtraTrees classifier achieved the highest accuracy (98.24%) and AUC (98.24%).
• Random Forest also demonstrated strong performance with 98.03% accuracy and AUC.
• The proposed ensemble approach outperformed existing state-of-the-art stroke prediction methods.

Conclusions:

• Ensemble machine learning, particularly the ExtraTrees classifier, offers a highly effective method for stroke prediction.
• This approach shows significant potential for improving early stroke detection and preventive healthcare strategies.