Chinese Noun Phrase Recognition Base on Word Co-occurrence Directed Graph
This paper proposes a recognition method for Chinese Noun Phrase based on word co-occurrence directed graph. An input document is firstly scanned in which noun word string is retrieved. Atomic word table and word co occurrence directed graph is then generated according to the word strings. A search is performed on the graph to find the longest paths with priority weight satisfying certain criteria. The word strings corresponding to the paths are considered as noun phrases. As dimensionality reduction is applied, the scale of the word co-occurrence directed graph is reduced significantly, and thus the efficiency of the algorithm is improved. Experimental results demonstrate that the precision of noun phrase recognition reaches 95.4％.
Atomic word Noun phrase Word co-occurrence directed graph Knowledge acquisition Natural Language Processing
Xing-Lin Liu Qi-Lun Zheng Qian-Li Ma
School of Computer Science and Engineering South China Univ. of Tech. Guangzhou, China School of Com School of Computer Science and Engineering South China Univ. of Tech. Guangzhou, China
国际会议
成都
英文
1788-1792
2010-12-17（万方平台首次上网日期，不代表论文的发表时间）