Automatic recognition of Chinese organization name based on conditional random fields

Person, location and organization have been always mentioned as a bottleneck of a Named Entity Recognition (NER) system. Automatic recognition of Chinese organization name is the most difficult problem in NER tasks. This paper presents a new approach of Chinese organization name recognition based on cascaded conditional random fields. In the proposed approach, we first recognize the person name and location name before recognizing organization. The model structure has been designed with the cascade way, the result then is passed to the high model and suppose the decision of high model for recognition of the complicated organization names. And we proposed the new feature to realize this task. We evaluate our approach on large-scale corpus with open test method using People's Daily (January, 1998), Chinese ORG recalling rate achieves 88.78% and the precision rate is 82.35%. The evaluation results show that our approach based on cascaded conditional random fields significantly outperforms previous approaches.