Related Experiment Video
Updated: Nov 8, 2025
07:35
Selecting Multiple Biomarker Subsets with Similarly Effective Binary Classification Performances
Published on: October 11, 2018
7.8K
R-HEFS: Rough set based heterogeneous ensemble feature selection method for medical data classification.
Rubul Kumar Bania1, Anindya Halder1
1Department of Computer Application, North-Eastern Hill University, Tura Campus, Tura 794002, Meghalaya, India.
Artificial Intelligence in Medicine
|April 20, 2021
Summary
A new ensemble feature selection method using rough set theory (RST) effectively reduces dimensions in medical datasets. This approach improves classification accuracy for disease diagnosis by selecting relevant, non-redundant features.
Area of Science:
- Machine Learning
- Data Mining
- Bioinformatics
Background:
- Feature selection is crucial for dimensionality reduction in large datasets.
- Ensemble feature selection (EFS) enhances learning algorithm performance and robustness.
- Existing methods may not optimally balance feature relevance and redundancy.
Purpose of the Study:
- To propose a novel Rough Set Theory (RST) based heterogeneous Ensemble Feature Selection (R-HEFS) method.
- To effectively select highly relevant and less redundant features from medical datasets.
- To improve classification accuracy and reduce computational complexity in disease diagnosis.
Main Methods:
- Developed R-HEFS integrating five RST-based filter methods.
- Utilized feature-class, feature-feature rough dependency, and feature-significance measures.
- Applied k-nearest neighbor (kNN) imputation and RST-based discretization for data preprocessing.
Main Results:
- R-HEFS effectively removed non-relevant and redundant features during aggregation.
- The method demonstrated improved classification accuracy on 7 out of 10 benchmark medical datasets.
- Evaluated using Naïve Bayes, Random Forest, SVM, and AdaBoost classifiers.
Conclusions:
- The proposed R-HEFS method is effective for dimensionality reduction in large medical datasets.
- R-HEFS enhances classification accuracy, aiding physicians in disease diagnosis with reduced complexity.
- This RST-based EFS approach offers a robust solution for medical data analysis.
