基于遗传算法的网络信息过滤系统中的特征选择

Feature Selection in Information Filtering System Based on GA

随着Internet在全世界的应用和普及,网络跟人们的生活越来越紧密,人们在使用网络的同时,不可避免地接触到大量的不良信息,为了过滤网络信息,出现了网络信息过滤。本文首先采用向量空间模型表示文本,给出了一个基于遗传算法的文本特征选择方法。同时为了克服遗传算法所存在着的收敛速度慢、易陷入局部最优等缺陷,文中采用了候选特征子集,并增加了一个扩展操作。最后,通过实验验证了该方法的有效性。

In order to keep people away from bad information from Intemet, the information filtering appeared. This paper gave a feature selection of text method based Genetic Algorithm, adopting Vector Space Model to present the texts, and an expanding operation to improve the GA＇s efficiency. An experiment was also presented to prove the useful of the method.