A Multi-Agent Strategy For Both English And Chinese Text Chunking

The traditional English text chunking approach identifies phrases by using only one model and phrases with the same types of features. It has been shown that the limitations of using only one model are that: the use of the same types of features is not suitable for all phrases, and data sparseness may also result. In this paper, a Multi-Agent strategy is proposed and applied in the identification of English phrases. And then, this strategy is rapidly transplanted to Chinese text chunking. This strategy puts phrases into agents according to their sensitive features and identifies different phrases in parallel, where the main features are: first, easy and quick communication between phrases; second, avoidance of data sparseness. Through testing on public corpus (English) and Chinese Penn Treebank (Chinese), F score of English chunking achieves to 95.70% and that of Chinese chunking is 95.25%. These results are higher than the best results that have been reported.