The Technology Research of The Semantic Text Classification
Guixian Xu1, Lirong Qiu1
COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(12C) 794-800
1College of Information Engineering, Minzu University of China
Semantic text classification is to classify the text according to the concepts of semantic relation. It can improve the performance of classification. This paper provides an efficient and accurate method of semantic text classification. First, the classification ontology is constructed by using the concepts extracted from Hownet. Second, Text is represented by semantic vector and general vector space. Then the semantic similarity calculation method is proposed among concepts. The similarity of concepts is calculated based on it. At last, semantic text classification is conducted based on KNN. The comparison of semantic classification and traditional classification is studied. Experiments show that the text classification method based on semantic relation can improve the classification accuracy effectively. The research is meaningful in the application of text clustering, information retrieval, natural language processing and construction of high-quality Tibetan corpus.