Design And Implementation Of Named Entity Recognition Algorithm For Financial Field

Named entity recognition refers to the recognition of entities with specific significance in the text,including human names,place names,institutional names,proper nouns,etc.,which is a basic and key task in natural language processing.It plays an important role in the process of natural language processing technology becoming practical.In order to improve the recognition accuracy of named entities in the financial field,a named entity recognition algorithm based on bidirectional LSTM neural network and CRF is designed and implemented.The algorithm mainly completes the following work.First of all,the information background in the financial field and the information of common and unique entities in this field are deeply understood,the differences and their respective characteristics between the proprietary named entity and the universal named entity are grasped,and the main flow and network structure of the algorithm are determined.Secondly,the word vector training corpus is collected and the training word vector model is processed,and the word vector model with the best effect is obtained by adjusting the training parameters iteratively.Then,collect crawling network training The data set is tested,the bidirectional LSTM network and CRF state transition matrix are trained,and the model is updated iteratively.Finally,the accuracy of each functional module in the process of named entity recognition algorithm is tested,and the module is optimized.The named entity recognition algorithm based on bidirectional LSTM and CRF is designed and implemented in the BosonNLP public named entity recognition dataset,which includes 2000 test sentences and 8000 named entities,which is compared with the named entity recognition algorithm based on LSTM.The accuracy rate increased by 2percentage points,the recall rate by 4 percentage points,and the F1 value by 3.03 percentage points.