加权贝叶斯线性B细胞表位特征提取方法
An weighted linear B-cell epitope Bayes feature extraction

特征提取方法对线性B细胞表位预测起到非常重要的作用，但贝叶斯特征提取方法忽略了氨基酸之间的相互关系. 为了更准确地描述表位序列的关系，提出一种基于氨基酸对量表加权的贝叶斯特征提取方法，该方法对单个氨基酸在序列分布的基础上充分考虑了氨基酸之间的关系，并使用支持向量机作为分类器进行分类. 在El-Manzalawy，Saha数据集上的测试表明改进的贝叶斯特征提取方法. 相比传统的贝叶斯特征提取方法，提取精度有一定的提升.
Feature extraction method play a very important role in linear B-cell epitope prediction. The bayes feature extraction methods ignore the relationship between the amino acids. In order to describe the epitope sequence more accurately，proposed an weighted bayes feature extraction(WBFE) based on amino acid antigen scale，apart to the distribution of the individual amino acids in the sequence the method is fully taking into account the relationship between the amino acid，and using support vector machine as classifier for classification. The test on the the El-Manzalawy，Saha data set shows that，compared to the traditional the Bayes feature extraction method，using the proposed method predicted effects enhance