Research And Implementation Of Chinese Medical Named Entity Recognition For Electronic Medical Records

In recent years,with the popularization of electronic medical records and the deepening of medical information construction,medical information processing has become increasingly important.In the field of medical information processing,medical named entity recognition is an important task.Its purpose is to identify medical-related entities,such as diseases,drugs,and operations,from text,and provide basic data for subsequent medical information processing.However,due to the large number of Chinese medical terms and complex expressions,Chinese medical named entity recognition has always been a challenging task.In order to fully adapt to the characteristics of the Chinese medical field,this paper designs a Chinese medical named entity recognition algorithm based on radical enhancement and vocabulary enhancement and a nested named entity recognition algorithm based on multi-head annotation,and develops an electronic Chinese Medical Named Entity Recognition System for Medical Records.The specific research contents are as follows:(1)This paper proposes a Chinese medical named entity recognition algorithm TsERL based on radical enhancement and vocabulary enhancement.For the enhancement of radical information,this paper constructs a radical map in the global stage,which strengthens the semantic information of characters.For vocabulary enhancement,this paper builds a word graph at the sentence level,and the model can perceive entity boundaries through this sentence-level structure.Experiments show that the algorithm’s Chinese medical named entity recognition effect is better than multiple comparison models based on radical enhancement and vocabulary enhancement.(2)This paper proposes a nested named entity recognition algorithm LPMA based on multi-head annotation.In order to better locate the positional relationship between tokens in the annotation matrix,this paper adds additional global positional information when using token pair-based representation to make up for semantic loss.This paper also designs a negative sampling strategy for multi-head labeling methods to effectively filter low-quality negative samples.The experimental results show that on the nested named entity recognition dataset,the recognition effect of the algorithm is better than other comparison models based on multi-head annotation.(3)This paper designs and implements a Chinese medical named entity recognition system for electronic medical records.The system realizes functions such as user management,data management,entity recognition,and visual display of results.The entity recognition part combines the TsERL algorithm with the LPMA algorithm to realize the effective identification of medical entities.The system has good stability and response speed,and provides a friendly user interface.