Explainable AI-driven intelligent system for precision forecasting in cardiovascular disease.

Area of Science:

• Artificial Intelligence in Medicine
• Cardiovascular Disease Research
• Healthcare Informatics

Background:

• Cardiovascular diseases (CVDs) pose a significant global health challenge, complicated by the difficulty of early, accurate prediction.
• Traditional machine learning models for CVD prediction often function as "black boxes," limiting clinical trust and usability.
• Explainable Artificial Intelligence (XAI) offers a potential solution by providing transparency into AI decision-making processes.

Purpose of the Study:

• To introduce an intelligent forecasting system for cardiovascular events utilizing Explainable Artificial Intelligence (XAI).
• To address the limitations of traditional, opaque machine learning models in predicting cardiovascular diseases.
• To enhance the trustworthiness and usability of AI-driven predictions in clinical settings.

Main Methods:

• Developed an intelligent forecasting system integrating advanced machine learning algorithms with XAI.
• Utilized a comprehensive dataset of 308,737 patient records from Kaggle, including demographics, clinical measurements, and lifestyle factors.
• Applied XAI techniques to provide understandable explanations for AI-driven cardiovascular event predictions.

Main Results:

• The proposed XAI system achieved 91.94% accuracy in predicting cardiovascular events.
• The system demonstrated a reduced miss rate of 8.06% compared to previous methods.
• XAI integration enhanced the transparency and trustworthiness of AI predictions for healthcare professionals.

Conclusions:

• Explainable Artificial Intelligence (XAI) significantly enhances the transparency and reliability of cardiovascular disease prediction.
• The developed XAI system improves clinical decision-making, leading to better patient care and treatment delivery.
• XAI holds substantial potential to advance cardiovascular healthcare through increased trust and usability of AI tools.