Cardiovascular disease detection using a novel stack-based ensemble classifier with aggregation layer, DOWA operator,...

Area of Science:

• Cardiovascular Diseases
• Machine Learning
• Medical Diagnostics

Background:

• Cardiovascular diseases (CVD) represent a significant global health challenge, impacting quality of life and leading to premature mortality.
• Early detection and intervention are crucial for mitigating CVD severity, progression, and mortality.
• Machine learning (ML) offers a promising avenue for advancing early CVD detection capabilities.

Purpose of the Study:

• To develop and evaluate a novel stack-based ensemble classifier for improved cardiovascular disease detection.
• To enhance classification accuracy and reliability in identifying cardiovascular disease instances.

Main Methods:

• Feature transformation using Johnson transformation and normalization of feature distributions.
• A stack-based ensemble classifier incorporating an aggregation layer and the dependent ordered weighted averaging (DOWA) operator.
• Utilizing three diverse first-level classifiers and a linear support vector machine (SVM) meta-classifier for final classification.

Main Results:

• The proposed ensemble classifier achieved an overall accuracy of 94.05%, a 5% improvement over baseline methods.
• The system demonstrated a significant increase in the area under the receiver operating characteristic (ROC) curve (AUC), reaching 97.14%.
• The enhanced classifier showed robust performance in distinguishing between positive and negative cardiovascular disease instances.

Conclusions:

• The addition of an aggregation layer to the stacking classifier significantly boosts classification accuracy for cardiovascular disease detection.
• The proposed method exhibits superior performance and reliability compared to recent studies in CVD classification.
• The developed classifier holds potential for effective and robust early detection of cardiovascular diseases.