TQCPat: Tree Quantum Circuit Pattern-based Feature Engineering Model for Automated Arrhythmia Detection using PPG...

Area of Science:

• Biomedical Engineering
• Signal Processing
• Quantum Computing Applications

Background:

• Arrhythmia poses significant morbidity and mortality risks.
• Photoplethysmography (PPG) signals offer a non-invasive method for assessing blood flow and detecting cardiac irregularities.
• Developing accurate and cost-effective arrhythmia detection methods is crucial.

Purpose of the Study:

• To propose a novel, self-organized feature engineering model for arrhythmia detection.
• To leverage simple, cost-effective photoplethysmography (PPG) signals for enhanced diagnostic capabilities.
• To develop an accurate system for classifying different types of arrhythmias.

Main Methods:

• Feature extraction using discrete wavelet transform (MDWT) and a quantum-inspired Tree Quantum Circuit Pattern (TQCPat).
• Feature selection employing Chi-squared (Chi2) and neighborhood component analysis (NCA).
• Classification using k-nearest neighbors (kNN) and support vector machine (SVM) with information fusion.

Main Results:

• The TQCPat-based feature engineering model achieved a classification accuracy of 91.30%.
• The model was validated on a large dataset of 46,827 PPG signals.
• Six distinct classes of arrhythmias were classified with ten-fold cross-validation.

Conclusions:

• The proposed TQCPat model demonstrates high accuracy for arrhythmia classification using PPG signals.
• The model's efficacy suggests potential for broader clinical application.
• Further validation with larger databases and additional arrhythmia classes is recommended.