Chinese Proprietary Domain Text Named Entity Recognition

Abstract

Chinese named entity recognition(CNER) is a basic task of natural language processing, the purpose of which is to identify the boundary and type of entity existing in the text. In this paper, we first deeply analyze the classical Chinese named entity recognition algorithm from two perspectives. On the basis of statistical knowledge of the conditional random field(CRF) and bidirectional long short term memory(Bi-Lstm) recurrent neural network for Chinese named entity recognition task. In combination with the advantages of Bi-Lstm deep learning entity features, to make up for the shortcoming that Bi-Lstm cannot dynamically output the results at the current moment by using the last output results. We introduce the CRF algorithm after the BI-Lstm algorithm, which makes the model can obtain the context semantic relations and ultimately improve the effectiveness of the experiment.