节点文献
基于特征选择的网页分类方法研究
Study on web pages categorization based on feature selection
【摘要】 随着网络信息的迅猛发展,信息处理已经成为人们获取有用信息不可缺少的工具,文本自动分类系统是信息处理的重要研究方向。对文本分类关键技术中的特征选择算法进行了探讨,并结合网页特性,对特征权重算法及互信息算法进行了改进。实验结果证明,改进算法是可行的。
【Abstract】 With the rapid development of information networks,information processing has become an indispensable tool for obtaining useful information,the text automatic categorization systems is an important research direction of information processing.The feature selection algorithms in the automated text categorization technology are deeply analyzed,and then the algorithm of term weighting and the mutual information algorithm are improved in view of the construct character of the web text.At last,the experimental results show that,the improvement algorithm is feasible.
【关键词】 自动分类;
特征选择;
向量空间模型;
互信息;
准确率;
【Key words】 automatic categorization; feature selection; vector space mode; mutual information; precision;
【Key words】 automatic categorization; feature selection; vector space mode; mutual information; precision;
- 【文献出处】 计算机工程与设计 ,Computer Engineering and Design , 编辑部邮箱 ,2007年17期
- 【分类号】TP393.092
- 【被引频次】8
- 【下载频次】189