Hospital Outpatient Guidance System Based On Knowledge Graph

As the beginning of the entire medical process, guidance plays a crucial role in helping patients analyze their condition and make the correct department selection. However, a common issue in many hospitals is the excessive number of patients in outpatient clinics, limited guidance desks, and the presence of volunteers with high turnover and low specialization as guidance personnel. This often results in prolonged waiting times and significantly impacts the overall patient experience. To address these problems, a knowledge graph-based hospital outpatient guidance system with chat and inquiry functions is proposed. The system aims to resemble a question-answering robot. The research focuses on three aspects: firstly, utilizing the Neo4j database to store data and create a medical knowledge graph; secondly, employing the Bi-LSTM+CRF model, LR+GBDT model, and Bert+textCNN model to perform named entity recognition(NER), intent recognition(IR), and slot filling tasks respectively; finally, implementing the interactive interface design using HTML5, CSS, and JavaScript. Experimental results demonstrate that the system successfully achieves chat and inquiry functions when tested with various types of queries and intents. This reduces patients’ waiting time during hospital visits, providing them with a convenient, intelligent, and efficient medical experience. Additionally, this research can serve as a reference for the intelligent upgrading and digital transformation of outpatient medical services.