Research On Intelligent Question Answering System Based On Knowledge Graph And Application On Medical Scenario

With the improvement of people’s living standards and economic development,more and more attention is paid to the concept of health and people’s livelihood,and people’s demand for medical consultation is also increasing.Traditional search engines are also difficult to meet people’s growing need for knowledge acquisition,and they are often faced with the following challenges:1)The demand for medical consultation is large,and the accuracy of the traditional medical question answering system is insufficient.2)The technical threshold and industry threshold for building an intelligent question answering system are high,and it is difficult to obtain and label the training corpus.3)There is no intelligent question answering system that meets the depth of knowledge in the field of hypertension consultation.In view of the above problems and challenges,this paper designs and implements an intelligent medical question answering system based on knowledge graph.The main research contents are as follows:1)Design and implement an intelligent medical question answering system based on knowledge graph.Separate business and technology,reduce dependence on medical experts,lower the threshold of computer knowledge,and efficiently build an intelligent medical question answering system.Supports graph management,corpus annotation,and intent configuration.After the business personnel import the graph,configure the intent and mark the corpus,they can train an intelligent medical question answering system with professional knowledge with one click.2)Propose and implement a corpus construction scheme and a question answering scheme based on knowledge text.Design and implement a corpus construction tool based on entity and relationship splicing,construct training samples by replacing different entities and corresponding relationships in the ontology domain,introduce a contrast loss function,and improve the language model’s ability to match natural language questions and knowledge texts,to improve the accuracy and stability of the Q&A process.3)Professional knowledge was extracted based on professional literature in the field of hypertension.Combined with the intelligent consultation for the hypertension scene,the ontology domain and entity relationship are designed,the knowledge map of hypertension is constructed,and the intelligent medical question answering system in the field of hypertension knowledge is realized for this scene.