Feature Selection and Dwarf Mongoose Optimization Enabled Deep Learning for Heart Disease Detection.

Area of Science:

• Medical Data Analysis
• Machine Learning Applications
• Cardiovascular Health

Background:

• Heart disease remains a leading global cause of mortality.
• Effective heart disease prediction is crucial for timely medical intervention.
• Current data mining and machine learning methods show limitations due to insufficient test data.

Purpose of the Study:

• To enhance the efficacy of heart disease detection performance.
• To introduce a novel hybrid feature selection method for improved accuracy.
• To develop an optimized deep learning model for precise prediction.

Main Methods:

• Data preprocessing involved quantile normalization and missing data imputation.
• A hybrid feature selection approach using Congruence coefficient Kumar-Hassebrook similarity was employed.
• Heart disease prediction was performed using SqueezeNet optimized by the dwarf mongoose optimization algorithm (DMOA).

Main Results:

• The DMOA-SqueezeNet model achieved high performance metrics.
• Maximum accuracy reached 0.925.
• Sensitivity and specificity were recorded at 0.926 and 0.918, respectively.

Conclusions:

• The proposed hybrid feature selection and DMOA-SqueezeNet model significantly improve heart disease prediction accuracy.
• This approach addresses limitations of existing methods by enhancing data quality and model optimization.
• The findings suggest a promising direction for developing more effective tools in cardiovascular health management.