Exploration of Part-of-Speech Tagging in The History of the Ming Dynasty Using Graph Models Based on Conditional-Random-Fields
Zhu Xia · Fudan xuebao. Ziran Kexue ban · 2014
Natural language processing is one of the most important fields in artificial intelligence.It has made great development along with the progressing of computer information process technology.But there is scarce study focused on Classical Chinese language processing.Here,we discussed Part-of-Speech tagging problem in The History of the Ming Dynasty.We applied three graph models of conditional random fields on the annalistic style history—The History of the Ming Dynasty.The results showed that complete graph model and nested graph model performed much better than no edge model in Part-of-Speech tagging of The History of the Ming Dynasty,and word segmentation helped improving precision and recall rate of its Part-of-Speech tagging.Part-of Speech tagging of unlisted words in testing set was rather low for conditional random fields based methods regardless of word segmentation.