A Text Categorization Method Based on Features Clustering

Abstract:

Choosing Features of a text is an important part of text categorization. Its result can affect the quality and efficiency of the text categorizer. Since there are usually thousands of features of a text, it always needs to reduce the dimension of the feature space. Considering the semantic relationship among words, a new text categorization method based on features clustering is proposed in this paper. This method first uses word segmentation to split texts into words, then, remove stop words and words with low information, and then calculate the distribution of words in these texts to construct a matrix of co-occurrence words. After that, cluster algorithms are employed to reduce the dimension of the feature space. Finally some experiments are carried out on two corpuses using several text categorization algorithms. The results demonstrate that this new method can not only improve the precision and recall of text categorization, but also increase the efficiency.

• Periodical: Advanced Materials Research (Volumes 532-533)
• Pages: 1090-1094
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.532-533.1090
• Online since: June 2012
• Authors: Zhi Bin Feng, Ping Jian Zhang, Juan Juan Zhao
• Keywords: Feature Selection, Features Clustering, Matrix of Co-Occurrence Words, Text Categorization

References

[1] Jiawei Han, Micheline Kamber, data mining concept and technology 2nd condition, China Machine Press, (2006).

[2] Guoa, B., Damper, R., Gunna, S.R., Nelsona, J.: A fast separability-based feature selection method for high-dimensional remotely sensed image classification. Pattern Recognition 41, 1653–1662 (2007).

DOI: 10.1016/j.patcog.2007.11.007

[3] Turney P.D. Learning Algorithms for Key phrase Extraction. Information Retrieval, 2, pp.303-336(2000).

[4] Institute of Computing Technology Chinese Academy of Science. Institute of Computing Technology, Chinese Lexical Analysis System. http: /ictclas. org/index. html.

[5] Xiaoxu Zhong. Research on key words selection using hierarchical clustering algorithm. Computer knowledge and technology, 2009, pp: 1483-1484.

[6] Salton G. A vector space model for automatic indexing. Communications of the ACM, 1975, 18: 613-620.

DOI: 10.1145/361219.361220

[7] ESTER M, KRIEGEL H, SANDER J. A density-based algorithm for discovering clusters in large spatial databases with noise. Proc of the 1996 2nd Int' l Conf on Knowledge Discovery and Data Mining. Portland : AAAI Press , 1996 : 226-231.

[8] Salton G and Buckley C. Term-weighting approaches in automatic text retrieval. Information Processing and Management, 1988, 24(5): 513-523.

DOI: 10.1016/0306-4573(88)90021-0

[9] LI Juanzi, FAN Qi'na, ZHANG Kuo, Keyword extraction based on tf/idf for Chinese news document, Wuhan University Journal of Natural Sciences. VOl. 12 NO. 5 2007 917-921.

DOI: 10.1007/s11859-007-0038-4

[10] The Sogou corpus, The R&D Center of SOHU, http: /www. sogou. com/labs/dl/t. html.

[11] Yang Yiming. An evaluation of statistical approaches to text categorization. Journal of Information Retrieval, 1999, 1(1/2): 67-88.

[12] Yang Yiming, Liu Xin. A re-examination of text categorization methods. In: Proceedings of ACM SIGIR Conference on Research and Development in Information Retrieval(SIGIR'99), 1999: 42-49.

DOI: 10.1145/312624.312647

[13] Osuna E, Freund R, Girosi F. An improved training algorithm for support vector machines. In: Proceeding of IEEE NNSP, 1997: 276-285.

DOI: 10.1109/nnsp.1997.622408