摘要

Domain specific terms and relations are very important resources in the relevant applications of NLP and knowledge management. In this paper, we propose a new relation extraction method combining pointwise mutual information and linguistics information. We utilize search engine to scrawl the latest pages containing a title and content, then extract entity relation from them. Our approach can be used in any domain corpus. We have done experiments in "SPORT", "EDUCATION" and "AUTOMOBILE" domains and then checked the domain specific relations extracted by our method and evaluated our approach, the results of the experiments show that our approach is more efficiency than baseline system.