Construction Of Diabetes Knowledge Map Based On Chinese Natural Language Processing

With the improvement of people's living standards,the incidence of diabetes is increasing year by year.Sustained hyperglycemia and long-term metabolic disorders can cause damage to systemic organs,especially the eyes,kidneys,cardiovascular and nervous systems,as well as dysfunction and exhaustion.However,the number of grassroots doctors in China is not enough,and the professional level is uneven,moreover there are many types of diabetes.Only by correctly understanding the types of diabetes can we help the people to effectively prevent and treat diabetes.In recent years,natural language processing technology has developed rapidly.This technology can be used to extract medical entities and their relationships from medical texts.Using the extracted knowledge to build a medical knowledge map,and successfully transform unstructured data into structured data.The medical knowledge map can assist medical staff in the diagnosis and treatment of diseases,at the same time it can better popularize medical knowledge to the people,and accelerate the development of the medical industry.At present,the use of natural language processing technology to build a knowledge map is a hot topic in academic research,and it is also widely used in various industries.This paper uses Chinese natural language processing technology to extract knowledge from the diabetes medical literature to construct a map of diabetes knowledge.These medical documents contain a large amount of medical information,which is of great significance for the prevention,diagnosis and treatment of diabetes,Because these medical documents are unstructured,if manual extraction is used,it will take a lot of manpower and material resources.How to extract the knowledge efficiently and accurately in the documents is the focus of text researchIn this paper,through the investigation and research on the construction process of medical knowledge map,the construction of diabetes knowledge grape based on Chinese natural language processing which is mainly divided into three parts: named entityrecognition,relationship extraction and knowledge map construction.In the named entity recognition part,this paper proposes a BERT-BiLSTM-CRF named entity recognition model.Based on the traditional BiLSTM-CRF model,this model incorporates the BERT word embedding model,which better integrates the context of the article and fully considers issues such as polysemy.In the relationship extraction part,a new joint relationship extraction model is constructed in this paper.Two subtasks of named entity recognition and relationship extraction share the coding layer.At the same time,during the training process,the sum of the loss functions of the two subtasks is optimized as the final loss function,which enhances the interactivity between the two subtasks.The knowledge map construction part constructs a diabetes knowledge map based on the Neo4 j map database,details the Neo4 j map database and the process of building a knowledge map,and conducts a simple analysis of the diabetes knowledge map.The successful construction of this knowledge map can be further applied to medical recommendation systems,medical auxiliary diagnosis and treatment systems,etc.,which are important to help to the prevention,diagnosis,treatment,and rehabilitation management of diabetic.