基于SVM分类机的一种DNA序列判别方法

A Classification Method of DNA Sequence Based on Support Vector Machine

针对DNA序列类别的分属问题,提出采用支持向量机(Support Vector Machine,SVM)的方法进行分类。根据SVM分类器的要求建立特征属性空间,首先由每个DNA中4个碱基的含量得到4个特征属性,然后在此空间中扩充DNA序列长度的属性,最后根据SVM分类器对已知的DNA分类样本做训练得到分类超平面。利用此超平面检测所要分类的DNA序列,实验结果表明这种方法具有很好的分类精度。

A new classification method of DNA sequence based on SVM was presented. The feature attribute space was established according to requirement of SVC. At first, four feature attributes were built by content of DNA＇s four bases. By increasing length attribute of DNA sequence in the space to extent the feature attribute space. Finally, the classification of hyperplane was obtained on the basis of available samples training by using SVC in the feature attribute space. The DNA sequence to be classified was verified by the hyperplane. The results show that the classification method accuracy is very good.