Network-based Support Vector Machine for Classifying Microarray Samples
Author Information
Author(s): Zhu Yanni, Shen Xiaotong, Pan Wei
Primary Institution: University of Minnesota
Hypothesis
Can a network-based support vector machine improve the classification of microarray samples by incorporating gene network information?
Conclusion
The proposed network-based support vector machine can effectively identify clinically relevant genes and improve classification accuracy in microarray data.
Supporting Evidence
- The network-based SVM identified more clinically relevant genes compared to standard methods.
- Simulation studies showed improved predictive performance in both low- and high-dimensional settings.
- The method was applied successfully to real microarray data related to Parkinson's disease and breast cancer.
Takeaway
This study created a new method to help computers better understand gene data by looking at how genes work together, which can help doctors find important genes related to diseases.
Methodology
The study used a network-based support vector machine that incorporates gene network information to classify microarray samples.
Limitations
The classification error rates may be biased due to the double use of data for training/tuning and testing.
Participant Demographics
105 patients, including 50 cases of Parkinson's disease and 55 controls.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website