基于遗传算法和支持向量机的肿瘤分子分类

提出了一种基于遗传算法(GA)和支持向量机(SVM)的用于肿瘤分子分类和特征基因选择的新方法。该方法针对基因表达数据样本少维数高的特点,先根据基因的散乱度滤掉大量分类无关基因,而后使用相关性分析去除分类冗余基因,得到一个候选基因子集,用遗传算法搜索候选特征基因空间,发现在支持向量机分类器上具有好的分类性能的且含基因个数较少的特征子集。把这种GA/SVM方法应用到结肠癌和急性白血病基因表达谱,能选出多个取得较高分类精度的较小基因子集,实验结果表明了该方法的有效性。

Tumor Molecular Classification Based on Genetic Algorithms and Support Vector Machines

A novel approach for the gene selection problem in gene expression and the cancer classification is presented.The method combines support vector machines(SVM) and genetic algorithms(GA).Firstly,the mess level of each gene is used as the criterion for filtering the irrelevant genes for classification.Then a correlation-based approach is adopted to remove the redundancy of the filtered subset and a candidate subset is generated.A genetic algorithm is developed to evolve gene subsets and the fitness is evaluated by SVM classifiers and their size.If the method is used on the two well-known colon and leukemia dataset,small subsets can be selected and the classification accuracy improved.Finally,the result demonstrates the effectiveness of the method.