A novel Chinese organization name extraction approach using CCRFs in micro-blog

Abstract

With the rapid growth of internet-scale, social media services (such as micro-blog) have gradually become a popular platform for the dissemination of information, which provides new research direction for Named Entity Recognition(NER). However, the informal structure and incorrect grammar in micro-blog and the complexity of Chinese expression make a serious challenge for NER task. Automatic identification of Chinese organization name in many NLP task is a difficult problem. To solve this problem, this paper presents a novel method for Chinese organization name recognition based on cascaded conditional random fields (CCRFs) where the identification task is divided into two subtasks. Firstly, the person name and location name in the micro-blog texts are first recognized by a Conditional random fields (CRFs) model. The result then is passed to another CRFs model and supports the decision of this model for identification of the complicated organization names. The experiment results show that the proposed method delivers better performance than existing methods.