Gene Selection Algorithms for Microarray Data
Author Information
Author(s): Tang E Ke, Suganthan PN, Yao Xin
Primary Institution: Nanyang Technological University, Singapore
Hypothesis
Can new gene selection algorithms improve classification accuracy in microarray data analysis?
Conclusion
The proposed gene selection approaches can provide gene subsets leading to more accurate classification results, while their computational complexity is comparable to the existing methods.
Supporting Evidence
- The GLGS algorithm generally achieves the lowest external B.632+ error among the compared methods.
- LOOCSFS outperforms marginal filter methods and other gene selection algorithms in terms of generalization error.
- The computational time of GLGS does not significantly increase with the number of genes selected.
Takeaway
Scientists created new methods to pick important genes from a large set, which helps in diagnosing diseases more accurately.
Methodology
The study proposed two algorithms, LOOCSFS and GLGS, for gene selection based on leave-one-out cross-validation error of LS-SVM, tested on two microarray datasets.
Limitations
The study does not address the optimal number of genes to select for all datasets.
Participant Demographics
60 patients with hepatocellular carcinoma and 50 glioma samples.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website