Research On Chinese Named Entity Recognition Method Based On GRU-CRF

Chinese named entity recognition is one of the basic tasks in the field of Chinese natural language processing.It is the basic of tasks such as machine translation and information extraction in the comprehensive application of Chinese natural language processing,which directly affects the performance of subsequent tasks of natural language processing.With the growth of network text resources,new knowledge is efficiently discovered and acquired from the network,and knowledge mining and analysis are more conveniently carried out through text.Chinese named entity recognition plays a major role.Deep learning-based Chinese named entity feature extraction algorithm is studied and achieved to better identify the physical features of Chinese named entity,and to improve the system's ability for Chinese named entity features,which has been a important content for named entity recognition research based on deep learning.The named entity recognition based on deep learning is regarded as the main research content.Firstly,the solution to the identification of named entities at home and abroad is studied.Then,relevant neural network methods based on deep learning for solving the named entity recognition in recent years are analyzed.The Chinese named entity feature extraction algorithms based on recurrent neural network,long short-term memory,gated recurrent unit are deeply studied.Finally,a new solution to the problem of Chinese named entity identification is proposed.A neural network model combining gated recurrent unit(GRU)and conditional random field(CRF)is proposed for Chinese named entity recognition.The application of the model in the Chinese named entity recognition task is deeply studied,and the feature extraction effect of Chinese named entity in the model training process is optimized.The problem of lack of Chinese corpus in the experiment is solved,so that a large number of Chinese corpora can be effectively handled with a small amount of manual annotation and Chinese named entity with large corpus of Chinese data can be effectively identified in a short time.Finally,the experimental comparison and analysis of the neural network model based on GRU and CRF shows that the neural network annotation model has good Chinese named entity recognition effect in Chinese named entity recognition task,especially in terms of name and organization name,the recognition effect is superior to other neural network models.