Abstract
The molecular characterization of different tumor types using gene expression profiling is expected to uncover fundamental aspects related to cancer diagnosis and drug discovery. There is, therefore, a need for reliable, accurate sample classification tools, as well as methods for efficient identification of genes informative for the class discrimination. We propose a method based on Support Vector Machine (SVM) ensembles, trained within a boosting framework. The approach allows sequential training of classifiers on different data subsets, their aggregate yielding results superior to single SVM. Results from binary and multiclass classification experiments performed on several data sets are presented.
Original language | English (US) |
---|---|
Pages (from-to) | 63-70 |
Number of pages | 8 |
Journal | Cancer Genomics and Proteomics |
Volume | 3 |
Issue number | 1 |
State | Published - 2006 |
Externally published | Yes |
Keywords
- Boosting
- Classification
- Ensemble learning
- Feature selection
- Gene expression
- Support vector machines
ASJC Scopus subject areas
- Biochemistry
- Molecular Biology
- Genetics
- Cancer Research