Research On Breast Tumor Aided Diagnosis Model Based On Knowledge Graph

Breast cancer has always been a disease that plagues many women and the most common cause of death from cancer.It has a great impact on women’s physical and mental health.Although there are a large number of clinical data on breast tumors in relevant domestic hospitals,these clinical data are relatively scattered and complex,and it is difficult to reflect the logic and relevance of the data.Knowledge graph technology can extract professional knowledge from massive data and has become the most important form of knowledge representation in the era of big data.The application of knowledge graph technology in the medical field can effectively integrate and integrate scattered medical knowledge.At present,mainstream medical knowledge graphs are mostly based on relevant medical websites and related documents,which can be used for commonsense medical knowledge services,but they are difficult to apply to specific diseaseassisted diagnosis.Therefore,this article mainly studies how to integrate the Internet’s Shanghai volume,scattered medical data and the clinical data of breast tumors continuously generated within the hospital system,construct a knowledge map of breast tumors,and provide knowledge services for breast tumors.The breast tumor assisted diagnosis model based on the knowledge map proposed in this paper is mainly divided into three parts: knowledge modeling of breast tumors,knowledge extraction of breast tumors,and knowledge services for assisted diagnosis of breast tumors.The clinical guide,as a special guide and standard manual used by clinical medical technicians,contains clinical diagnostic standards and clinical treatment plans.According to the clinical diagnosis and treatment guidelines for breast tumors,this article first builds a model layer.Secondly,a large amount of clinical data will be generated during the diagnosis and treatment of patients in the hospital and stored in the electronic medical record(EMR).We use EMR as the data source to extract patient-centric knowledge of breast tumors.Then,based on the medical website data,we use neural network technology to perform entity recognition and relationship extraction to obtain knowledge about drugs and symptoms related to breast tumors.Finally,build a knowledge map centered on patient information,covering drugs,symptoms,examinations and other information.In order to complement the lack of relationships in the knowledge map of breast tumors,we proposed a multi-step reasoning model to complete the completeness of the knowledge map of breast tumors.And based on the completed knowledge map of breast tumors,we constructed an auxiliary diagnosis model.The final experimental results show that our multi-step reasoning model improves the relationship prediction of implicit connections by 6.3%,complements the implicit relationships between entities and optimizes the knowledge map of breast tumors.On this basis,we took the diagnosis of benign and malignant breast tumors as a classification task,and proposed the Bert-SVM-agg model to accurately classify the benign and malignant breast tumors.By comparing with the basic classification model,the overall classification accuracy has been increased by 4.2%,verifying the effectiveness of the model...