一种基于模糊遗传算法的最优特征子集优化方法研究  

Optimization of A Subset of Features Based on Fuzzy Genetic Algorithm

作　　者：解英杰[1,2] 朱振方[2] 

机构地区：[1]山东师范大学信息科学与工程学院,250014 [2]山东技师学院信科系,250200

出　　处：《信息技术与信息化》2008年第6期35-37,共3页Information Technology and Informatization

摘　　要：针对基于内容的网络信息过滤中存在的特征维数过高影响分类过滤效果问题,运用遗传算法进行特征选择,通过遗传操作搜索最优解。并且引入个体寿命概念用于实施种群更新,同时根据种群进化情况动态调整遗传操作算子,从而解决遗传算法训练过程中种群以及操作算子的单一性带来局部最优问题。实践证明,该方法能够实现特征子集空间规模的压缩,有效提高过滤效果。To the problem of feature dimension too high to effect filtering in Content - based network information filtering, using genetic algorithm to feature selection, and search the optimal solution through genetic manipulation. And introduce the concept of individual life to update population. At the same time, adjustment genetic operator dynamically for the population evolution, so as to solve single of population and operator in genetic algorithm in the course of training. Practice has proved that this method can be achieved subset of the features of the scale of the compressed space, and effectively improved the filtering effect.

关 键 词：模糊理论 遗传算法 特征选择 信息过滤 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]