Improved Joint Kazakh POS Tagging and Chunking

摘要：

　　This paper describes a mixing model of joint POS tagging and chunking for Kazakh where partial optimal solution provide feature information for joint model.A improved beam-search algorithm use dynamic beam instead of unified beam to obtain search space of small-but-excellent during both training and decoding phases of the model.Moreover we can statistical induction the information of chunk to disambiguation of multi-category words and experiment shows the precision is improved from 81.6%to 87.7%by information of chunk.

关键词： mixing model joint model dynamic beam multi-category words

作者: Hao Wu Gulila Altenbek

作者单位: College of Information Science and Engineering,Xinjiang University,Urumqi,China;The Base of Kazakh and Kirghiz Language of National Language Resource Monitoring and Research Centre Minority Languages,Urumqi,China

会议类型: 国内会议

会议名称: 第十五届全国计算语言学学术会议(CCL2016)暨第四届基于自然标注大数据的自然语言处理国际学术研讨会(NLP-NABD-2016)

会议地点: 烟台

会议语种:英文

页码: 1-10

在线出版日期: 2016-10-14（万方平台首次上网日期，不代表论文的发表时间）