TFIDF算法研究综述  

Study of TFIDF algorithm

作　　者：施聪莺[1] 徐朝军[1] 杨晓江[1] 

机构地区：[1]南京师范大学教育技术系,南京210097

出　　处：《计算机应用》2009年第B06期167-170,180,共5页journal of Computer Applications

摘　　要：文本分类中特征项权重的赋予对于分类效果有较大的影响,TFIDF算法是权重计算的重要算法之一。在回顾TFIDF算法发展历史的基础上,考察了其固有缺陷,总结诸多学者对其的改进方法,并对TFIDF算法新的应用领域进行了概括,并通过实验验证相关改进算法,为读者更好地应用TFIDF算法提供参考。In text categorization, the weight of term has great impact on the classification results. Term Frequency and Inverse Documentation Frequency （TFIDF） is one of the key algorithms of term weighting. This paper reviewed the development of the TFIDF algorithm, studied its inherent defects, and summarized some scholars＇ improvements to it. Meanwhile, the survey generalized its new application fields. To verify their effects on the classification results, the author carried out some experiments on the ameliorative algorithms, hoping to provide some reference to readers.

关 键 词：TFIDF 文本分类 VSM 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]