论文部分内容阅读
This paper presents a new improved term frequency/inverse document frequency (TF-IDF) approach which uses confidence, support and characteristic words to enhance the recall and precision of text classification. Synonyms defined by a lexicon are processed