Cancer Gene Extraction Based on Stepwise Regression

摘要

With the expansion of the gene expression profile database, in the case of as little as possible to lose information or to retain the most critical information, gene extraction has become a main direction for the scholars. This paper excludes 1561 irrelevant genes through the definition of weighted distance firstly, and then removes 252 redundant genes by Pearson's correlation coefficient. Finally by comparing the two methods, stepwise regression after clustering and only stepwise analysis, we obtain the best combination of 8 genes.