ExF-SVM: Exhaustive feature selection with support vector machine algorithm for brain stroke prediction

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Machine Learning for Disease Prediction

Background:

• Brain stroke prediction is critical for patient outcomes but challenged by AI's "black box" nature.
• Explainable AI (XAI) aims to bridge the gap between AI potential and clinical trust by enhancing model interpretability.
• Timely and accurate brain stroke prediction is essential to prevent severe patient harm and improve treatment efficacy.

Purpose of the Study:

• To propose a novel feature selection technique for identifying crucial characteristics in brain stroke prediction.
• To develop and assess an efficient brain stroke risk detection model using explainable AI.
• To enhance the accuracy and reliability of brain stroke prediction models in clinical settings.

Main Methods:

• Development and evaluation of an Exhaustive Feature Selection with Support Vector Machine (ExF-SVM) algorithm.
• Utilizing feature selection to determine the most impactful characteristics for brain stroke risk.
• Assessing model performance using metrics such as Receiver Operating Characteristics (ROC) curve, sensitivity, specificity, and F1-Score.

Main Results:

• The proposed ExF-SVM algorithm demonstrated improved classification accuracy by 4-14% compared to existing models.
• The F1 score saw a significant enhancement of 5-15% with the implemented methodology.
• The results highlight the effectiveness of the novel feature selection technique in improving brain stroke prediction.

Conclusions:

• The developed ExF-SVM algorithm offers a more interpretable and accurate approach to brain stroke prediction.
• This explainable AI model can enhance clinical decision-making and patient care in stroke management.
• The findings suggest significant contributions and ramifications for healthcare through advanced AI applications.