Approaches to domain adaptive chinese segmentation model

CORC > 北京大学 > 信息科学技术学院

	Approaches to domain adaptive chinese segmentation model
	Han, Dong-Xu ; Chang, Bao-Bao
刊名	jisuanji xuebaochinese journal of computers
	2015
DOI	10.3724/SP.J.1016.2015.00272
英文摘要	Character-based tagging method is currently one of effective methods in Chinese Word Segmentation (CWS). Constrained by domain and size of the training corpus, this method doesn't work well in domain adaptability, affecting its use in practical application. This paper puts forward using chi-square statistics and boundary entropy to enhance the segmentation method in handling the Out-Of-Vocabulary words. Combined with self-training and co-training strategies, we further improve the performance of domain adaptability in CWS. Experiments show that with the use of these proposed methods, the domain adaptability of CWS is effectively improved. ?, 2015, Science Press. All right reserved.; EI; 0; 2; 272-281; 38
语种	英语
内容类型	期刊论文
源URL	[http://ir.pku.edu.cn/handle/20.500.11897/329400]
专题	信息科学技术学院
推荐引用方式 GB/T 7714	Han, Dong-Xu,Chang, Bao-Bao. Approaches to domain adaptive chinese segmentation model[J]. jisuanji xuebaochinese journal of computers,2015.
APA	Han, Dong-Xu,&Chang, Bao-Bao.(2015).Approaches to domain adaptive chinese segmentation model.jisuanji xuebaochinese journal of computers.
MLA	Han, Dong-Xu,et al."Approaches to domain adaptive chinese segmentation model".jisuanji xuebaochinese journal of computers (2015).