Chaotic gradient based optimization with fuzzy temporal optimized CNN for heart failure prediction

Area of Science:

• Cardiology
• Artificial Intelligence
• Biomedical Engineering

Background:

• Heart failure is a major cause of premature death, particularly in sedentary individuals.
• Early and accurate detection is crucial for preventing heart failure progression.
• Existing prediction systems often lack early accuracy and are time-consuming.

Purpose of the Study:

• To develop an advanced heart failure detection model for improved early and accurate identification.
• To enhance feature selection and classification accuracy in heart failure prediction.
• To address the limitations of current heart failure detection systems.

Main Methods:

• Proposed a novel Chaotic Gradient-Based Optimizer (CGBO) combining chaotic maps and Gradient-Based Optimizer (GBO).
• Introduced the Fuzzy Temporal Optimized Convolutional Neural Network (FTOCNN) classifier, integrating CGBO and fuzzy temporal rules.
• Evaluated the model using the UCI heart dataset and Electronic Health Records (EHRs), employing statistical measures, classification metrics, and tenfold cross-validation.

Main Results:

• The CGBO method significantly improved feature selection for heart failure risk.
• The FTOCNN classifier achieved 94% accuracy in heart failure detection using EHR data.
• The proposed model demonstrated superior performance compared to various Machine Learning (ML) and Deep Learning (DL) classifiers.

Conclusions:

• The developed CGBO-FTOCNN model provides a significant advancement in early and accurate heart failure detection.
• This approach enhances the reliability of heart failure prediction systems.
• The findings suggest a promising new direction for managing heart failure through advanced computational methods.