A Knowledge Graph Based System On The Accurate Doctors’ Recommendation

The filtration of professional knowledge is one of the hotspots of the modern information style.The choice of doctors in the medical field is one of the typical problems,which is that the information of hundreds of doctors within the scope of registration can not accurately and selfFeatures to match,and it’s more difficult to make a best choice.Aiming at this problem,this paper makes a doctor recommendation system based on the knowledge graph and the similarity degree of the instance,which has the background of Guangdong Provincial People’s Hospital.The doctor recommendation system take the domain knowledge base as the professional knowledge source,which links up the doctors,patients,patients,disease,etc,to suit for the business characteristic,and take the recommendation based on the knowledge recommendation.Based on these two points,this paper introduces the relevant technical knowledge,and analyzes the essence of those technology and the doctor recommendation’s common feature,then by the in-depth understanding of the doctor recommendation,point out the limitation of the relevant technical knowledge.After analyzing the technical and business problems,this paper begins to propose how to solve the problem of doctor’s professional information recommendation through the combination of technology and innovation.In this paper content,we begin to study the specific application details of OWL,which is the leading edge technology ontology language.Then,paper give a set of demand-driven,from coarse to fine domain knowledge graph construction scheme,the first is the knowledge expression and then describe step by step the extraction of coarse-grained demand performance data based on Web crawler,the fine-grain enrichment based on electronic semi-structured medical documents,the optimization based on the knowledge engineers and domain experts.The evaluation method of the ontology take the realization of the whole construction scheme to finish.After the introduction and construction of the knowledge,the author starts to build the model of the information in the knowledge graph based on the idea of case-based recommendation,and puts forward two modeling concepts of "Doctor Record" and "Patient Record".The semantic similarity between the cosine similarity and the case-based similarity is proposed to improve the similarity calculation based on semantic,and the concept of Matching Degree is put forward in this part.The entire modeling and calculation are just right on the abstract refinement in solving the problem of measurement.At the last,the design of the whole doctor recommendation system is introduced,and the analytic and presentation module and doctor recommendation module are the major part.The whole system design scheme uses a lot of theory of software engineering,especially highlighting the ability of control and reversal of knowledge.Finally,we evaluate key point of the whole doctor recommendation system.