With the advancement of information technology,the Internet has become an increasingly important role in social life,and more and more patients choose to obtain medical information through the Internet.The medical knowledge data obtained through the traditional search engine services of the Internet is scattered,and the amount of it is large,it is difficult for patients to obtain the medical knowledge they need.At the same time,the rapid development of natural language processing technology provides a solid foundation for dialogue systems to become a new way of human-computer interaction.In the medical field,the intelligent dialogue system can fully understand the patient’s questions,provide concise answers to the patient’s questions,and improve service efficiency.Therefore,taking the dialogue system in the medical field as the research direction,this paper designs and implements an intelligent medical consultation system.The intelligent medical consultation system designed in this paper includes a database module,a consultation function module,and a user interaction module.In the database module,the medical knowledge data acquisition,data cleaning,and data extraction are completed.Then the data is stored in the Neo4j database according to the three classifications of named entities,inter-entity relationships,and entity attributes designed in this paper.In the Consulting module,this article completes the Intent Recognition task and the Named Entity Recognition task.In the intent recognition task,this paper uses the enriched and fused Chinese medical intent dataset to train a variety of intent recognition models.including the BERT intent recognition model,RNN intent recognition model,and Text-CNN intent recognition model and selects the Text-CNN intent recognition model with the best recognition effect in the experiment to provide services for the system.In the named entity recognition task,this paper uses the Chinese medical question answering named entity recognition dataset to train a variety of named entity recognition models,and selects the BERT-BiLSTM-CRF named entity recognition model with the best recognition effect in the experiment to provide services for the system.In the user interaction module,this paper realizes the interaction with the user in the form of designing a web page dialogue.Different from the existing medical dialogue system using pattern matching technology or end-to-end dialogue model,this paper designs and implements an intelligent medical consultation system using named entity recognition technology and intent recognition technology.After testing,the system can answer a variety of medical consultation questions,achieve the design goals of linking patients and professional medical knowledge,improve the efficiency of medical services and reduce the cost of some medical services,and enrich the research results of the medical dialogue system. |