论文部分内容阅读
针对基因微阵列数据高维小样本问题及可能存在的非线性结构问题,本文采用拉普拉斯特征映射(LE)方法对基因微阵列数据进行降维处理,再构造支持向量机(SVM)进行分类识别。文中对采用该方法与主成分分析线性降维方法的分类结果进行了分析对比,结果表明结合拉普拉斯非线性降维的支持向量机分类模型具有对基因数据更好的分类处理能力。
In order to solve the problem of high dimensional small sample of gene microarray data and possible nonlinear structure problems, this paper uses Laplace tracing mapping (LE) to reduce the dimension of gene microarray data and reconstruct support vector machine (SVM) Category identification. In this paper, the classification results of the linear dimensionality reduction method using this method and the principal component analysis are compared and analyzed. The results show that the SVM classification model combined with Laplace nonlinear dimensionality reduction has better classification ability for gene data.