Heart Disease Prediction Based on the Embedded Feature Selection Method and Deep Neural Network.

Area of Science:

• Cardiology
• Artificial Intelligence
• Machine Learning

Background:

• Heart disease poses a significant global health threat due to its high prevalence and mortality rates.
• Early prediction of heart disease using routine physical indicators is crucial for timely diagnosis and intervention.
• Manual analysis of large datasets for heart disease prediction is time-consuming and challenging.

Purpose of the Study:

• To develop an accurate and efficient model for predicting heart disease using readily available physical indicators.
• To leverage machine learning to overcome the limitations of manual data analysis in cardiovascular risk assessment.
• To identify key indicators significantly associated with heart disease for improved predictive accuracy.

Main Methods:

• A novel heart disease prediction algorithm combining embedded feature selection and deep neural networks was developed.
• Linear Support Vector Classification (LinearSVC) with L1 norm regularization was employed for feature selection.
• A deep neural network with He initializer weights was constructed for disease prediction.
• The model was evaluated on a heart disease dataset using metrics such as accuracy, recall, precision, F1-score, and AUC.

Main Results:

• The proposed model achieved high performance metrics: 98.56% accuracy, 99.35% recall, 97.84% precision, and a 0.983 F1-score.
• The average Area Under the Curve (AUC) score reached 0.983, indicating excellent predictive capability.
• The embedded feature selection effectively identified significant indicators for heart disease prediction.

Conclusions:

• The developed model demonstrates high efficiency and reliability in predicting heart disease.
• The integration of embedded feature selection and deep neural networks offers a promising approach for early cardiovascular risk assessment.
• This method can aid clinicians in making informed decisions and improving patient outcomes.