中文摘要：

建立了四类基于基因表达的分类器,用以将87名妇女的子宫内膜样本分成癌症患者和非癌症患者.首先利用信噪比过滤掉无关基因,然后利用主成分分析降低样本维数,再针对这四类分类器随机取75个样本作为训练样本,其余的12个样本作为测试样本,实验结果表明这四类分类器适合子宫内膜癌的分类.最后采用留一交叉验证作为评判标准,通过比较,说明5BP-ELMAN分类器是一类更适合子宫内膜癌分类的有效的肿瘤分类器.

英文摘要：

In the paper, four kinds of classifiers based on gene expression are built to classify 87 women＇ endometrial samples into one class with the endometrial cancer and the other class with cancer-free. Firstly, use SNRs to filter irrelevant genes and then principal component analysis method to decrease the dimensions. Randomly take 75 samples as trained samples and other 12 samples as tested samples for these classifiers. The results are these four classifiers are suitable for classification of endometrial cancer. Finally use leave-one-out cross validation as evaluation criteria. By comparison, the results are shown that 5BP-ELMAN classifier is an effective tumor classifier more suitably for classification of the endometrial cancer.