【摘要】 结合了基因表达数据类内和类间表达差异的信息,提出一种新的基因选择算法,利用它选择出来的特征基因表达作为支持向量机的输入特征向量,对四个常用数据集进行分类,结果表明,该方法可以显著提高分类精度,同时通过对选取出来的特征基因在相关信号通路上的分析,表明该方法能够得到更多的肿瘤相关基因,具有很强的鲁棒性和很高的精确度.更多还原

【Abstract】 A novel gene selection method based on tumor gene expression data was proposed.It incorporated the within-class and between-class variations of the gene expression values to select significant genes.It was evaluated on four publicly available tumor gene expression datasets,using leave-one-out cross-validation based on SVM classifier.The performance was measured by the classification accuracies and the top ranked genes are discussed and annotated in special disease pathways.The experimental results show that the proposed gene selection method is effective and robust.The selected genes by our proposed method in general give more accurate classification results and the top ranked genes are biologically significant.更多还原