Hybrid metaheuristic optimization for detecting and diagnosing noncommunicable diseases

Area of Science:

• Computational Biology
• Health Informatics
• Data Science in Healthcare

Background:

• Healthcare faces challenges in early Non-Communicable Diseases (NCDs) detection and management.
• The COVID-19 pandemic highlighted the need for advanced tools for NCD prediction and treatment, especially for at-risk populations.

Purpose of the Study:

• To propose a comprehensive framework integrating data mining, feature selection, and meta-heuristic optimization for NCDs.
• To develop and evaluate novel hybrid algorithms for efficient and accurate NCD classification.

Main Methods:

• Development of the Hierarchical Genetic Multiple Reduct Selection Algorithm (H-GMRA) for minimal feature set identification.
• Implementation of Customized Function-based Particle Swarm Optimization with Rough Set Theory for NCD Feature Selection (CPSO-RST-NFS).
• Extensive experimentation on diverse NCD datasets to validate the framework's performance.

Main Results:

• H-GMRA demonstrated superior performance over traditional methods in identifying feature subsets with high dependency ratios.
• CPSO-RST-NFS achieved improved efficiency and accuracy by combining meta-heuristic optimization with feature selection.
• The framework successfully selected informative features and enhanced classification accuracy across various NCD datasets.

Conclusions:

• The proposed framework significantly advances early detection and management of NCDs.
• This research bridges the gap between computational efficiency and disease classification accuracy, offering better patient outcomes.
• The findings provide valuable insights for healthcare practitioners and data analysts in NCD research and clinical prediction.