摘要
力求解决困扰传统进化树构造中只能生成二叉树、精度低和Tie Tree的问题。采用自组织神经网络对序列进行分类,生成进化树过程中允许扩展当前非叶子节点,并通过设置适当的参数优化进化树的分层。使用此方法,获得了精度更高的多叉进化树,表明基于神经网络的方法对解决进化树构造中的问题是有效的。
In this paper, the neural network of self-organizing map is used to overcome the tie tree drawback, which is a serious problem in traditional phylogenetic tree reconstruction, and improve the preciseness of phylogenetic tree. The multifurcating tree is also reconstructed to replace the role of traditional binary tree, so that the system evolution be displayed more precisely. By giving proper parameters, the arrangement and sequence cluster can be optimized. The experiments indicate that this method is effective.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2006年第B07期27-31,共5页
Journal of Harbin Engineering University
基金
北京市自然科学基金资助项目(4052005).