H-DSAE: a hybrid technique to recognize heart disease

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Diagnostics

Background:

• Heart disease is a leading global cause of mortality, with risk factors including high cholesterol and blood pressure.
• Conventional diagnostic methods can be time-consuming and prone to human error, leading to potential inaccuracies.
• Machine learning offers a promising approach to improve diagnostic accuracy and inform clinical decision-making for cardiac conditions.

Purpose of the Study:

• To develop an advanced machine learning model for accurate heart disease classification.
• To improve upon conventional diagnostic methods by leveraging deep learning techniques.
• To reduce the likelihood of cardiac events through more informed and precise diagnoses.

Main Methods:

• An ensemble deep learning method, H-DSAE, was employed for heart disease classification.
• The H-DSAE technique integrated Deep Belief Network (DBN), Support Vector Machine (SVM), and Stacked Auto-Encoder (SAE).
• The model was designed to extract complex representations from heart images for classification.

Main Results:

• The H-DSAE technique achieved a high classification accuracy of 99.2%.
• The model demonstrated strong performance with a sensitivity of 97.5%, F-measure of 98.5%, and precision of 98.4%.
• The approach successfully extracted relevant heart image features for diagnostic purposes.

Conclusions:

• The H-DSAE deep learning technique significantly enhances the accuracy of heart disease classification.
• This advanced method offers a more efficient and reliable alternative to conventional diagnostic approaches.
• Future work will focus on further refining classification algorithms and feature extraction for improved system efficiency.