Classification of cardiac abnormalities using heart rate signals

Area of Science:

• Cardiology
• Biomedical Engineering
• Computational Neuroscience

Background:

• Heart rate (HR) is a non-stationary signal with variations indicating current or impending cardiac diseases.
• Analyzing long-term HR data for abnormalities is time-consuming and strenuous.
• Heart rate variation (HRV) measurement is a non-invasive tool for assessing autonomic nervous system function.

Purpose of the Study:

• To develop and evaluate computer-based analytical tools for in-depth study and classification of day-long HR data.
• To improve the efficiency and accuracy of cardiac rhythm classification for diagnostic purposes.

Main Methods:

• Classification of cardiac rhythms using an artificial neural network (ANN).
• Integration of fuzzy relationships with ANN for enhanced data analysis.
• Utilizing instantaneous heart rate against time measurements.

Main Results:

• The study achieved a high level of efficacy in classifying cardiac rhythms.
• The developed tools demonstrated an accuracy level of 80-85% in classification tasks.
• Computer-based analysis proved useful for diagnostics over day-long intervals.

Conclusions:

• Artificial neural networks combined with fuzzy logic offer a highly effective method for cardiac rhythm classification.
• The computational approach significantly aids in the non-invasive assessment of cardiac conditions.
• This methodology enhances diagnostic capabilities for cardiovascular health monitoring.