Research On Chinese Text Oriented Named Entity Intelligent Recognition Model

Named entity recognition aims to identify and mark entities with specific meanings in text.It is a key technology to further extract entity relationships and mine other potential information in natural language processing.At present,the methods based on machine learning and deep learning have been widely used in the research of named entity recognition,but most learning models use feature extraction based on word and character level.The word preprocessing of this kind of model often ignores the context semantic information of the target word and can’t realize polysemy.In addition,the loss of semantic information and limited training data sets greatly limit the improvement of model performance and generalization ability.In order to solve the above problems and improve the efficiency of named entity recognition technology in Chinese text,this paper constructs a multi task BERT-BiLSTM-AM-CRF intelligent processing model.The main research work is as follows:(1)The BERT-BiLSTM-CRF single task model is constructed.The dynamic word vector combined with context information is extracted by BERT.After further training by BiLSTM module,the result is input into CRF layer for decoding.CRF classifies and extracts the obtained observation annotation sequence to obtain the final result.In the single task experiment part,a control experiment is designed for each module of the model.Through the comparison of experimental data,it is verified that the Bert model improves the overall efficiency of the model.(2)Based on the single task model,a multi task learning model is constructed.After extracting the word vector through BERT and inputting it into BiLSTM for feature learning,through the attention mechanism network,the model can learn together on two Chinese data sets,and finally get the training results through CRF layer constraints.Compared with many previous single task models,the F1 score of this multi task model in MASR and People’s Daily data sets has been significantly improved(0.55%and 3.41%),which demonstrates the effectiveness of multi task learning on Chinese named entity tasks.(3)The training time of multi task model is too long.In order to ensure the accuracy of the model and futher shorten the training time of the model.This paper takes the transformer,a feature extractor of Bert,as the core,improves and implements a parallel Chinese named entity recognition model.The experimental results show that the improved transformer model in this paper achieves better results on MSRA data set than the original transformer model.Compared with the multi task learning model,the training speed is significantly improved,which demonstrates the effectiveness of the model.