Chinese Text Classification Based On LDA and KSVM
With the rapid development of information technologyand social networking, the amount of generated text data has increased enormously. As one of the crucial technologies for information organization and management, text classification has become much more significant in the area of machine learning and natural language processing. According to this paper, we present a text classification system. First, we apply LDA topic model to express the text instead of Boolean model or vector space model. Then, we choose KSVM which combines SVM with KNN as the classification algorithm. Finally, we choose documents with large amount of Chinese news for experiments. Compared with normal language models, these experimental data shows that our system gets higher classification accuracy.
Machine Learning Text Classification LDA KSVM KNN SVM-KNN
Congwei Liang Yong Liu Haiqing Du
Beijing University of Posts and Telecommunications, Beijing, China
国际会议
重庆
英文
379-383
2015-12-18（万方平台首次上网日期，不代表论文的发表时间）