English Location Entity Recognition Based on Conditional Random Fields

Location entity detection is one task of automatic content extraction 2008 (ACE2008), which defines five types of entities. In this paper, the conditional random fields model is used to recognize the English location by selecting different characteristics. We select the features of English location dictionary, preposition prefix to improve the experimental results. Through calculating the recall, precision and F-value, we found that they are great impact of the quality by the characteristics. The experimental results show that, to a certain extent, more characteristics are selected, better result will be got.