Identification of Heart Sounds with an Interpretable Evolving Fuzzy Neural Network

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Informatics

Background:

• Heart disease is a leading global cause of mortality.
• Intelligent techniques can improve pattern identification in cardiac diseases.
• Early detection of heart murmurs is crucial for identifying potential pathological conditions.

Purpose of the Study:

• To develop and evaluate an evolving hybrid intelligent technique for heart murmur detection.
• To extract knowledge from heart noise behavior data using fuzzy linguistic rules.
• To assess the model's performance against state-of-the-art methods in heart disease detection.

Main Methods:

• Utilized an evolving hybrid intelligent technique for data analysis.
• Applied fuzzy linguistic rules for knowledge extraction from heart noise data.
• Compared the model's accuracy with existing state-of-the-art approaches for heart disease detection.

Main Results:

• The proposed hybrid model achieved 90.75% accuracy in detecting heart murmurs.
• Demonstrated high assertiveness in identifying heart murmurs.
• Successfully extracted knowledge from the analyzed data through an intelligent approach.

Conclusions:

• The evolving hybrid model is effective for accurate heart murmur detection.
• The model facilitates knowledge extraction, enhancing the understanding of cardiac data.
• This approach supports medical decision-making in cardiology.