针对统计方法不能从语义理解的角度进行文本分类的问题，提出了利用概念层次网络概念知识进行文本分类的方法，包括两部分：依据概念进行特征选取以及根据类别关联度分类. 在特征选取时，通过计算概念与类别的区分度挖掘出类别核心概念，并采用类别核心概念对特征项进行精选. 依据类别核心概念相关的类别语义信息，提出了文档与类别关联度的计算方法，并根据类别关联度来判断文本类别. 实验表明，该方法可有效降低特征空间维数，在提高分类效率的同时保证了分类效果，F1值略有提高. 与SVM、KNN和Bayes分类器对比，当特征项数目较少时，该方法的F1值明显高于其他3种方法，综合分类效果与SVM相当，优于KNN和Bayes.
To achieve semantic understanding, this paper proposes a method for text categorization based on concept-knowledge in the hierarchical network of concepts (HNC). The method includes two parts: feature selection using concepts and text categorization according to category relatedness degree. In this paper, category key concepts are explored by computing discrimination degree of concepts, and used to further reduce dimensionality of the feature space. Based on the category semantic information consisting of category key concepts and relatedness weights, the method of computing relatedness degrees between documents and categories is proposed. The category relatedness degree of document is used as a measure for text categorization. Experiments show that the proposed method can effectively reduce dimensionality of feature space, increase efficiency and ensure effectiveness of text categorization. Compared with SVM, KNN and Bayes, this method is the best in terms of F1 values at higher feature reduction levels. In terms of overall performance, the method is almost equivalent to SVM, and better than KNN and Bayes.
[1] YANG Yiming, LIU Xin. A re-examination of text categorization methods [C]//Proceedings of the 22nd Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, California, United States, 1999: 42-49.
[2] YANG Y, PEDERSEN JO. A comparative study on feature selection in text categorization [C]//Machine Learning-International Workshop Then Conference, San Francisco, USA, 1997: 412-420.
[3] OGURA H, AMANO H, KONDO M. Feature selection with a measure of deviations from Poisson in text categorization [J]. Expert Systems with Applications, 2009, 36(3): 6826-6832.
[4] SHANG Wenqian, HUANG Houkuan, ZHU Haibin, LIN Yongmin, QU Youli, WANG Zhihai. A novel feature selection algorithm for text categorization [J]. Expert Systems with Applications, 2007, 33(1): 1-5.
[5] KUMAR MA, GOPAL M. A comparison study on multiple binary-class SVM methods for unilabel text categorization [J]. Pattern Recognition Letters, 2010, 31(11): 1437-1444.
[6] Manabu TORIIA M, Lanlan YINB L, Thang NGUYENA T, Chand T. MAZUMDARA C T, Hongfang LIU H F, David M. HARTLEYA D M, Noele P. NELSONA N P. An exploratory study of a text classification framework for internet-based surveillance of emerging epidemics [J]. International Journal of Medical Informatics, 2011, 80(1): 56-66.
[7] 张孝飞，黄河燕. 一种采用聚类技术改进的KNN文本分类方法[J]. 模式识别与人工智能，2009, 22(6): 936-940.
ZHANG Xiaofei, HUANG Heyan. An improved KNN text categorization algorithm by adopting cluster technology [J].Pattern Recognition and Artificial Intelligence, 2009, 22(6): 936-940. (in Chinese)
[8] 李荣陆，胡运发. 基于密度的kNN文本分类器训练样本裁剪方法 [J]. 计算机研究与发展，2004, (04): 539-545.
LI Ronglu, HU Yunfa. A density-based method for reducing the amount of training data in kNN text classification [J]. Journal of Computer Research and Development, 2004, 41(4): 539-545. (in Chinese)
[9] WU G, CHANG E Y. KBA: kernel boundary alignment considering imbalanced data distribution [J]. IEEE Transactions on Knowledge and Data Engineering, 2005, 17(6): 786-795.
[10] LIU Xuying, WU Jianxin, ZHOU Zhihua. Exploratory under-sampling for class-imbalance learning [C]//Sixth International Conference on Data Mining, HongKong, China, 2006: 965-969.
[11] SUN A, LIM EP, LIU Y. On strategies for imbalanced text classification using SVM: a comparative study [J]. Decision Support Systems, 2009, 48(1): 191-201.
[12] 孙海霞，钱庆，成颖. 基于本体的语义相似度计算方法研究综述 [J]. 现代图书情报技术，2010, 9(1): 51-56.
SUN Haixia, QIAN Qing, CHENG Ying. Review of ontology-based semantic similarity measuring [J]. New Technology of Library and Information Service, 2010, 9(1): 51-56. (in Chinese)
[13] BAI Rujiang, WANG Xiaoyue, LIAO Junhua. Extract semantic information from wordnet to improve text classification performance [J]. Advances in Computer Science and Information Technology, 2010, 6059/2010: 409-420.
