Design And Implementation Of Knowledge Graph In The Field Of Medical Ultrasound

Medical ultrasound data can provide a scientific basis for the diagnosis and treatment of patients’ symptoms,and assist doctors in making medical decisions;in medical research,a large amount of ultrasound data provides researchers with basic research conditions.Researchers need to inquire about ultrasound knowledge and need to select data that meets the research conditions.If only through manual screening,advanced retrieval cannot be achieved;at the same time,in terms of quality control,the field of machine learning is now more and more widely used in medicine,and many task machines can Complete,how to control the quality of manual diagnosis based on the machine-generated results is an important subject of scientific research.This article is oriented to the field of medical ultrasound and provides a knowledge graph system that can solve the above problems.The main purpose of this paper is to study the construction of the knowledge graph in the field of ultrasound.Through the structured processing of the medical ultrasound unstructured text,the association relationship between entities is obtained,and the existing structured knowledge is combed to obtain the ultrasound knowledge graph.Research provides supporting information and facilitates scientific research for researchers.The main work is as follows:(1)Perform named entity recognition on ultrasound data.This paper proposes a named entity recognition model for ultrasound data,introduces Bert on the baseline model BiLSTM-CRF,and designs experiments to solve the problem of entity recognition for unstructured data.(2)Extract entity relationships from ultrasound data.This paper introduces the segmentation pooling and attention mechanism,adopts the PCNN-Attention model and joint rule matching method to classify the relationship of the entity pairs,and verifies the extraction effect through experiments.(3)Developed a scientific research assistance system based on the ultrasound knowledge graph.This paper constructs an ultrasound knowledge graph,stores the basic information and correlations of ultrasound in the graph database,and develops an auxiliary scientific research system based on the graph.After investigating the application scenarios,the system is aimed at researchers in the field of medical ultrasound.The development work includes the realization of functions such as data retrieval,data quality control,and related knowledge query.Among them,in the retrieval function module,the UER framework is used to realize the annotation of the input sentence,and the Neo4j graph database and Elasticsearch are used to realize the data filtering.And with the help of data visualization,a convenient and efficient scientific research support system is finally realized.