基于协同进化遗传算法的主题信息采集研究  

Research on Topic-specific Web Resource Based on Co-evolution Genetic Algorithms

作　　者：刘光洁[1] 李忠范[2] 李民[1] 杨鑫[1] 

机构地区：[1]长春师范学院计算机学院,吉林长春 130032 [2]吉林大学数学学院,吉林长春 130021

出　　处：《情报科学》2008年第10期1531-1534,共4页Information Science

摘　　要：利用改进的协同进化遗传算法,使全局搜索与局部搜索、全局收敛性与收敛速度有机地统一起来,并将之应用到主题信息采集的研究中,实例结果表明具有较好的分类预测能力。This paper improves coordinating operation in co-evolution genetic algorithms,a model based on global coordination and local evolution as well as improved genetic algorithms are proposed.A local evolution method based on neighbor field is adopted for enhancing capability of local search and accelerating convergent speed. At the same time, an improved multi-population co-evolution algorithm is used for the diversity of individuals in populations. The method is used to topic-specific web resource, and experimental results show this algorithm has good predictive ability.

关 键 词：遗传算法 协同进化 主题信息 收敛 URL 

分 类 号：G203[文化科学—传播学]