Research And Implementation Of Dialog Data Construction And Dialog Generation Based On Medical Knowledge

At present,in the context of the COVID-19 epidemic sweeping the world,online medical services play an important role in it,and online medical consultation services play an important role.In medical consultation,it is necessary to understand the dialogue and diagnose and treat patients.In an actual medical dialogue scenario,the medical dialogue between doctors and patients contains diseases,symptoms,drugs,tests and other information that is extremely important,and is an important basis for building smart medicine,which affects whether doctors can correctly diagnose and treat patients.In previous studies,medical data sets presented problems such as single label,insufficient number of labels,single training task,etc.,and there was no medical dialogue data set with multi task and multi label.In order to provide more effective medical consultation services,this topic provides a multi task and multi label medical dialogue data set,which is an important basis for building smart medicine.For this reason,this paper crawls the medical dialogue from the online Chunyu doctor website,organizes personnel and formulates labeling specifications,and constructs a multi task and multi label medical dialogue data set.This data set has played a constructive role in medical dialogue research.Secondly,for smart medicine,it is extremely important to help patients make auxiliary decisions.Based on the current generation of medical dialogue,there are problems such as insufficient medical entities to reply to traditional Chinese medicine and confusion in the relationship between medical entities,which reduce the credibility of the generation of medical dialogue and affect the judgment of patients.To solve these problems,this paper proposes an entity aware medical dialogue model that integrates knowledge maps.It uses the dialogue history to predict the medical entities that will respond in this round,and introduces medical knowledge maps to assist in generating medical dialogues,so that the credibility of medical dialogue generation can be improved and applied to the medical dialogue generation model.Finally,this paper applies the proposed model to the actual medical interrogation,designs and implements an intelligent dialogue system based on medical interrogation,and interfaces it with the We Chat applet.Through the functional demonstration of the medical interrogation dialogue system,it further verifies the feasibility of the entity aware medical dialogue model integrating knowledge map in real life applications.