Mining Knowledge of Respiratory Rate Quantification and Abnormal Pattern Prediction.

Area of Science:

• Computational intelligence
• Medical informatics
• Cardiovascular health

Background:

• Cardiovascular disease (CVD) is a leading global cause of mortality.
• Abnormal respiratory patterns are increasingly linked to CVD development and progression.
• Physician support for respiratory pattern evaluation is crucial.

Purpose of the Study:

• To develop a granular computing-based method for respiratory pattern evaluation.
• To differentiate between periodic-like, intermediate, and normal breathing patterns.
• To identify key predictors of abnormal respiration.

Main Methods:

• Application of granular computing principles: group decision-making, tri-way reasoning, and rough set-based analysis.
• Development of prediction models using signal attributes and anthropomorphic parameters.
• Validation of the methodology using k-nearest neighbor (k-NN) and uniform manifold approximation and projection (UMAP).

Main Results:

• Median classification accuracies exceeded 0.75 in many cases.
• Signal analysis parameters were found to be more significant than anthropomorphic features.
• High waist-to-hip ratio (WHR) and male sex were identified as predisposing factors for abnormal respiratory patterns.

Conclusions:

• The proposed methodology offers a valuable tool for physicians in respiratory pattern evaluation.
• Granular computing effectively aids in classifying breathing patterns relevant to CVD risk.
• Understanding risk factors like obesity and sex can improve CVD prevention strategies.