K-Mean Clustering Analysis and Its Applications to Classification of Tumor Gene
Feature gene selection of tumor classification is an important means to find the expression of tumor-specific genes. To study the tumor gene expression pattern, k-means clustering analysis method is considered. It is used for selecting the best genetic center, extracting scalar features and determining the corresponding gene label. The experimental results show that the correct rate of the classification results by this method is 87 %.
KeywordsGene expression profile Feature gene K-mean clustering
- 1.Golub TR, Slonim DK, Tamayo P et al (1999) Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 286(5439):531–537Google Scholar
- 3.Chun T, Aidong Z, Jian P (2003) Mining phenotypes and informative genes from gene expression data. Proc 9th ACM SIGKDD Int Conf Knowl Discov Data Min 4:655–660Google Scholar
- 4.Duda OR, Hart PE, Stork GD (2001) Pattern classification. second edition. Wiley, New York, 6(7):46–48Google Scholar
- 5.Theodoridis S, Koutroumbas K (2003) Pattern classification, vol 5 issue no 6, 2nd edn. Academic Press, New York pp 177-179Google Scholar
- 6.Wang SL, Wang J, Chen HW et al (2006) SVM-based tumor classification with gene expression data//international conference on advanced data mining and applications, vol 4093. Springer, Berlin, Heidelberg, pp 864–870Google Scholar