Related Experiment Video
Updated: Apr 5, 2026
03:37
Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers
Published on: March 1, 2024
1.5K
A Filter Feature Selection Method Based on MFA Score and Redundancy Excluding and It's Application to Tumor Gene
Jiangeng Li1,2, Lei Su3,4, Zenan Pang1
1Institute of Artificial Intelligence and Robotics, College of Electronic Information & Control Engineering, Beijing University of Technology, Beijing, 100124, China.
Interdisciplinary Sciences, Computational Life Sciences
|August 24, 2015
Summary
A new feature selection method, MFA score+, improves tumor gene expression analysis by reducing data redundancy. This technique enhances classification accuracy compared to existing methods like MFA score.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Feature selection is crucial for analyzing tumor gene expression data.
- Marginal Fisher Analysis score (MFA score) is a popular graph embedding-based method, outperforming Fisher score.
- Gene expression data often contains significant redundancy.
Purpose of the Study:
- To introduce a novel filter feature selection technique, MFA score+.
- To address redundancy in gene expression data using MFA score and redundancy exclusion.
- To evaluate the performance of MFA score+ against existing methods.
Main Methods:
- Developed MFA score+, a filter feature selection method based on MFA score and redundancy exclusion.
- Applied MFA score+ to an artificial dataset and eight tumor gene expression datasets.
- Utilized support vector machine (SVM) for sample classification.
Main Results:
- MFA score+ effectively selected important features from gene expression datasets.
- The proposed method achieved higher classification accuracy compared to MFA score, t test, and Fisher score.
- Demonstrated superior performance in tumor sample classification.
Conclusions:
- MFA score+ is an effective feature selection technique for tumor gene expression data.
- The method's ability to handle data redundancy leads to improved classification accuracy.
- MFA score+ offers a valuable advancement over existing feature selection approaches.
