A Feature Weight Algorithm for Text Classification Based on Class Information

Abstract:

TFIDF algorithm was used for feature weighting in text classification. But the result of classification was not very well because of lack of class information in feature weighting. The known class information in the training set was used to improve the traditional TFIDF feature weight algorithm. Class distinction ability and class description ability were introduced, respectively expressed by inverse class frequency and term frequency in class, document frequency in class. A new feature weight algorithm based on class information, TF_IDT, was proposed. Naïve Bayes classifier was used to test the algorithm. The precision, recall and F1 measure were significantly increased. Macro F1 measure raise by 6.46%. It was proved to be useful for improving text classification to use class information in feature weighting. In addition, the computational complexity of the proposed algorithm was lower and more suitable for use in fields of limited computing capability.

• Periodical: Advanced Materials Research (Volumes 756-759)
• Pages: 3419-3422
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.756-759.3419
• Online since: September 2013
• Authors: Yong Fei Li
• Keywords: Document Frequency in Class, Feature Weight, Inverse Class Frequency, Term Frequency in Class, Text Classification

References

[1] Franca D and Fabrizio S. Supervised Term Weighting for Automated Text Categorization, Proceedings of the 18th ACM Symposium on Applied Computing. Melbourne: ACM Press, 2003, pp: 784-788.

DOI: 10.1145/952532.952688

[2] Shi Cong-ying, Xu Zhao-jun and Yang Xiao-jiang. Comprehensive Research on TFIDF Algorithm, Computer Application, vol. 29, Jun. 2009, pp.167-170(in Chinese).

[3] Zhang Ai-hua, Jing Hong-fang and Wang Bin etc. Study on function of feature weighting factor in Text classification, Chinese Information Processing, vol. 24, Mar. 2010, pp.97-103(in Chinese).

[4] ZHANG Yu-fang, Peng Ming-shi and Lyu Jia. Improvement and application of TFIDF method in text classification, Computer Engineering, vol. 32, Oct. 2006, pp.76-78(in Chinese).

[5] Shen Zhi-bin and Bai Qing-yuan, improvement of feature weight algorithm in Text classification, Journal of Nanjing Normal University (Engineering and Technology), vol. 8, Apr. 2008, pp.95-98(in Chinese).

[6] Zhang Yu and Zhang De-xian, An improved feature weight algorithm, Computer Engineering, vol. 37, May. 2011, pp.210-212(in Chinese).

[7] Li Kai-qi, Diao Xing-chun and CAO Jian-jun, improved text feature weight algorithm based on information gain, Computer Engineering, vol. 37, Jan. 2011, pp.16-18, 21(in Chinese).

[8] Liu Ting, Qin Bing and Zhang Yu etc., Introduction to Information Retrieval System, Beijing: Mechanical Industry Press, 2008(in Chinese).