【摘要】 为了解决传统TF-IDF函数由于忽略特征项之间的联系带来的诸多问题,对TF-IDF函数在文本分类中的应用进行了研究.结合信息论相关知识,找出了类间分布度和类内分布度表征特征项之间的潜在关系,进而提出改进的TF-IDF函数用于文本分类.实验表明,改进后的TF-IDF函数是有效可行的,而且较好的弥补了传统方法所丢失的特征项之间的关联信息,提高了文本分类的准确率.更多还原

【Abstract】 In order to solve the problems brought by neglecting the relationship between different terms,the application of TF-IDF function in text categorization is researched in this paper.Combining the relative knowledge of information theory,find the latent relationship between the distribution information among classes and distribution information inside a class.And then an improved TF-IDF function is proposed which uses the distribution information among classes and distribution information inside a class.The experiment shows that the improved method is feasible and effective.In addition,it greatly improves the accuracy of text Categorization.更多还原