基于G-ICA的组织样本分类算法

Sample classification algorithm based on G-ICA

利用加入了分类指导信息的ICA(Guide Independent components analysis,G-ICA),在已知样本中提取隐藏在样本基因表达数据中与组织分类密切相关的各种表达模式,根据这些模式对未知组织样本进行分类。试验结果表明,该方法提高了组织样本的分类能力,其计算复杂度低、收敛快,具有较强的稳定性。

It uses the ICA that includes the information for classification（G-ICA） to absorb the modes lie in the gene expression data,then classify the samples without label upon these modes.The experiment shows that it can improve the ability of sample classification,in addition,it has low complexity and strong stability,and can get the target lastly.