Company Name Recognition Based on Hybrid Method

With the development of Internet, there is more and more information which needs to be progress based on language. Named Entity Recognition is a good method to promote models of machine learning to obtain the structure features and contextual constraint, thus understanding the whole sentence, while company name recognition is an important part among Named Entity Recognition. In this paper, a model based on BERT-BiLSTM-CRF and text match is put forward to realize company name recognition. The text match based on company name feature thesaurus can recognize abbreviations more efficiently, thus improving the accuracy and recall of company name recognition.