Research Of Doctor Recommendation Method Based On Community Detection

With the advent of the big data era,medical information has also been paid more and more attention by our country.In the chain of medical big data consisting four elements of medicine,disease,gene,and human,the element of human made up of patient and doctor is always at the core position.It goes without saying that doctors are very important in the whole doctor-patient activities,the communication and cooperation among doctors is the key point to smooth the progress of medical activities.This paper takes doctors as the object,constructs a cooperative network among doctors,and proposes the algorithm of SC-DP to conduct community detection of the constructed complex network.For doctors in the same community,this paper suggests them to form a medical team,at the same time,the KOL doctors will be searched based on the communities this paper have divided,finally the KOL doctors will be recommended.The purpose is to excavate the potential laws of the doctor's cooperative relationship,in order to guide the cooperation among doctors better,promote doctors to solve medical problems more accurately and efficiently,therefor,improve the level and quality of medical service much further.Firstly,this paper collects doctors' information and their corresponding medical institutions' information,moreover,the cooperative information of the academic papers between doctors and so on from the medical information platform through the distributed network crawler technology,fuse the information from different platforms by data preprocessing to make it fit the data specification required by the following research.Then take doctors as nodes,the number of the cooperative papers between doctors as the weight of the edge to construct the doctors' paper cooperative relationship network,the connected subgraphs are further constructed through the union-find algorithm and the graph is finally visualized by Gehpi.Secondly,based on the basic algorithms of modularity,spectral clustering,Chebyshev distance instead of Euclidean distance and Density Peaks clustering algorithm,this paper proposes an algorithm of SC-DP which has better effect and faster speed to conduct community detection on the dataset of the research,getting good results,Then combine the five degree segmentation theory to complete the whole process of community detection to obtain the final network and visualize it.Finally,this paper finds the KOL doctor based on the obtained communities,the criterion is that the doctors represented by the nodes with the largest degree are considered as KOL doctors and then they are recommended,namely,the KOL doctor recommendation method based on community detection.Through surfing and integrating the network data,the method is proved to be reliable,the recommended doctors have KOL characteristics and are able to play the KOL role.