英文摘要：
In the field of AI, intelligent dialogue systems have important research significance and application value. Intelligent dialogue system is a human-computer interaction system that communicates with people through natural language, and is defined as the entrance to various future services. However, there is relatively little research on intelligent dialogue systems in the field of problem diagnosis. In this article a task-based intelligent dialogue system applied in the field of problem diagnosis was designed and implemented, using the Pipeline Method architecture to design a dialogue process that includes natural language understanding, dialogue management, knowledge base, and natural language generation. And a novel knowledge base model and dialogue management method were proposed: a bipartite graph knowledge organization method based on problem symptoms and root causes, and a problem root cause dialogue strategy based on probability distribution. This system, as a solution for remote diagnosis, utilized artificial intelligence to connect key technical points in the field of medical general practice, forming a practical, usable, and engineering oriented general practice auxiliary diagnostic dialogue system. For the randomly selected case data of 23 common diseases from public publications, the recall rate of this system has achieved good results with the top four recall rates close to the diagnostic results of professional doctors, and this system achieves the effects of shortening response time, reducing costs, and improving customer satisfaction.
|