A Novel Feature Selection Method for High-Dimensional Biomedical Data Based on an Improved Binary Clonal Flower...

Area of Science:

• Biomedical data analysis
• Machine learning in healthcare
• Bioinformatics

Background:

• Rapid accumulation of biological and clinical data presents challenges in analysis.
• The "curse of dimensionality" hinders accurate disease diagnosis and at-risk patient assessment.
• Effective feature selection is crucial for improving the performance of biomedical data analysis.

Purpose of the Study:

• To propose a novel feature selection method to address the curse of dimensionality in biomedical data.
• To enhance the accuracy of disease classification using high-dimensional biological and clinical datasets.
• To improve the assessment of at-risk patients and diagnostic capabilities.

Main Methods:

• An improved binary clonal flower pollination algorithm for feature selection.
• Incorporation of an absolute balance group strategy and adaptive Gaussian mutation.
• Utilizing the K-Nearest Neighbors (KNN) classifier for evaluating classification accuracy.

Main Results:

• The proposed method effectively eliminates unnecessary features from high-dimensional biomedical datasets.
• Achieved high classification accuracy in experiments across six public datasets.
• Demonstrated superior performance compared to existing state-of-the-art methods.

Conclusions:

• The novel feature selection algorithm successfully overcomes the curse of dimensionality.
• The method significantly improves disease classification accuracy in biomedical applications.
• This approach offers a promising tool for enhanced patient risk assessment and diagnosis.