Title :
Ambiguity measure-based feature selection for text categorization
Author :
Yang, Jieming ; Liu, Zhiying
Author_Institution :
Coll. of Inf. Eng., Northeast Dianli Univ., Jilin, China
Abstract :
Feature selection is one of methods that reduce the size of the number of features in text categorization. In this paper, we proposed a feature selection method, which filtered some features that only rarely occur in one category and do not occur in other categories from the feature subset generated by ambiguity measure method. The experiments show that the proposed method can improve the performance in the context of special classifier and text corpus.
Keywords :
pattern classification; text analysis; ambiguity measure method; ambiguity measure-based feature selection; classifier; feature subset; text categorization; text corpus; Accuracy; Information filters; Mutual information; Niobium; Support vector machines; Text categorization;
Conference_Titel :
Intelligent Control and Information Processing (ICICIP), 2012 Third International Conference on
Conference_Location :
Dalian
Print_ISBN :
978-1-4577-2144-1
DOI :
10.1109/ICICIP.2012.6391414