摘要
由于串行化特征选择方法应用于中文海量文本数据集时时间效率较低,提出了一种用于特征选择的并行免疫克隆方法.该方法采用免疫克隆算法搜索特征,利用并行算法评价特征子集,即将种群中个体的亲和度计算并行在多个计算节点上同时进行,从而能较快地获得较具代表性的特征子集.实验结果表明,该方法是有效的.
Considering serial feature selection method is inefficient timely to be applied to Chinese massive text data sets, a feature selection method based on parallel immune cloning algorithm was presented. The method uses the immune cloning algorithm to select feature subset and calculates fitness of feature subsets in multiple computing nodes at the same time, so can acquire quickly the feature subsets which are more representative. The experimental results show that the method is effective.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2009年第12期1847-1851,1857,共6页
Journal of Shanghai Jiaotong University
基金
四川省科技计划项目(2008GZ0003)
四川省科技攻关项目(07GG006-019)
关键词
特征选择
免疫克隆算法
并行算法
亲和度
feature selection
immune cloning algorithm
parallel algorithm
fitness
