基于优化的文档频和Beam有哪些信誉好的足球投注网站的特征选择方法朱颢东.pdfVIP

36 11 V ol. 36 No. 11 2009 11 Computer Science Nov 2009 Beam ( 中国科学院成都计算机应用研究所 成都610041) ( 中国科学院研究生院 北京100039) 在文本分类中,特征空间的维数通常高达几万, 甚至远远超出训练样本的 个数, 这是一 十分普遍现象为 了提高文本挖掘算法的运行速度, 降低占用的内存空间, 过滤掉不相关或相关程度低的特征, 必须使用特征选择算法 首先给出了一 个基于最小词频的文档频方法, 然后把粗糙集引入进来并提出了一 个基于Beam 有哪些信誉好的足球投注网站的属性约简算法, 最后把该属性约简算法同基于最小词频的文档频方法结合起来, 提出了一 个综合的特征选择算法该算法首先利用 基于最小词频的文档频方法进行特征选择, 然后利用所提属性约简算法消除冗余, 从而获得较具代表性的特征子集 实验结果表明该算法是有效的 词频, 文档频, 粗糙集, Beam 有哪些信誉好的足球投注网站, 属性约简 TP301 A Feature Selection Method Based on Optimized Document Frequency and Beam Search ZH U H ao-dong ZH ONG Yong ( Chengdu Institute of Computer Application , Chinese Academ y of Sciences, Chengdu 610041, China) ( T he Graduate School of the Chinese Academy of Sciences, Beijing 100039, China) Abstract In text cat egorization, one problem is usually confronted w ith feature spaces containing 10, 000 dimensions and more, even exceeding the number of available training samples. In order to enhance the operating speed and reduce the memory space occupied and filter out irrelevant or low er degree of features, feature select ion algorithms must be used. In order to obtain more representat ive feature subset , it firstly presented document frequency method based on minimum word frequency , and then introduced rough sets and presented an algorithm of attribute reduct ion based on Beam search,finally , combined the attribute reduction algorithm wit h document frequency method based on minimum word frequency and proposed a comprehensive feature select ion algorithm. T he comp