The nature of statistical learning theory
The nature of statistical learning theory
Neural Networks for Pattern Recognition
Neural Networks for Pattern Recognition
Cancer gene search with data-mining and genetic algorithms
Computers in Biology and Medicine
Two-stage classification methods for microarray data
Expert Systems with Applications: An International Journal
Extracting gene regulation information for cancer classification
Pattern Recognition
Accurately learning from few examples with a polyhedral classifier
Computational Optimization and Applications
A study of cross-validation and bootstrap for accuracy estimation and model selection
IJCAI'95 Proceedings of the 14th international joint conference on Artificial intelligence - Volume 2
Hi-index | 0.00 |
The growing availability of biological measurements at the molecular level has recently enhanced the role of machine learning methods for effective early cancer diagnosis, prognosis and treatment. These measurements are represented by the expression levels of thousands of genes in normal and tumor sample tissues. In this paper we present a two-phase algorithm for gene expression data classification. In the first phase, a novel gene selection method based on mixed-integer optimization is applied with the aim of selecting a small subset of cancer marker genes. In the second phase, a binary polyhedral classifier is used in order to label gene expression data. Computational experiments performed on three benchmark datasets indicate the usefulness of the proposed framework which is capable of competitive performances with respect to the best classification accuracy so far achieved for each dataset. Moreover, the classification rules generated are based on very few genes which, in our computations, can be credited as the most influential genes for tumor differentiation.