Research Of Chinese Named Entity Recognition Based On Deep Learning

Named entity recognition is one of basic tasks in natural language processing,the improvement of named entity recognition accuracy for information retrieval,automatic question answering,relationship extraction has a lot of help.In recent years,with the generation of large amounts of data in all walks of life,accuracy and applicability of named entity recognition system are put forward new requirements.This paper mainly does work as follows:(1)Studied the named entity recognition based on statistical principle,points out the defects and disadvantages of statistical methods,and analyzed the newest method of named entity recognition by using deep learning.(2)On the basis of deep learning,improve the training way of word vector,put forward the weighted word vector.(3)Analyzed the problems existing in the current deep learning framework for naming entity recognition,and proposed an improved framework.Analyzed the reason of named entity recognition in many fields precision is not high,proposed the new framework that basing the sematic analysis and neural network;Analyzed the similarity of between named entity recognition and machine translation,proposed the idea that solving the named entity recognition by machine translation way,modified the model of machine translation to be appropriate for named entity recognition,and achieved good results.(4)The designed experiment to verify the effectiveness of the proposed the new framework and introduces the application of named entity recognition in intelligent question answering system.