Research On Medical Case Information Mass And Case Base Maintenance Considering With Labelled Attribute

Due to the coming of the big data era,and the continuous development and construction of medical informatization,it makes the information resources also showing the characteristics of explosive growth,and the increasing complexity and variability of patients' diseases,these have made it challenging for medical workers to make rational and accurate diagnosis and treatment decisions,therefore,the current research in the field of medical decision management faces two problems to be solved,firstly,in the face of a large number of decision information sources,how to quickly find more accurate,more reasonable,more appropriate similar cases and solutions,and provide support knowledge for decision making of medical staff.Secondly,under the current condition of changing environment of diagnosis and treatment,whether the final result of diagnosis and treatment decision can be made more scientific,reasonable and accurate by considering more external attribute factors.Based on these,this paper studies from two aspects,namely the research of improvement and application innovation of case-based reasoning process,as well as the research of case information aggregation and case base maintenance considering the labelled attribute information.At first,this paper expounds the knowledge requirement of medical staff for diagnosis and treatment decision under the current environment,and also introduces the improved case-based reasoning process as retrieval technology for historical experience and knowledge,and then further summarizes and analyzes the present research situation and existing shortages of case-based reasoning method,on this basis,a similar cases retrieval method based on the angle-distance is proposed,and the data set of coronary heart disease(CAD)in public data platform UCI is analyzed experimentally to verify the validity,rationality and accuracy of the proposed method in the management of chronic disease diagnosis and treatment.Finally,through compared the experimental results of the other classification methods in chronic disease management,it is proved that the proposed method is more applicable and effective.Second,in view of the increasingly complex and changeable condition of the patient's disease,and the constantly changing decision-making environment,this paper puts forward to consider the labelled attribute information outside the medical case,that is,the authority of the case source and the doctor's evaluation of the case,and then explores its effect on the results of the similar cases,the outcome of medical staff's final diagnosis and treatmentand the maintenance of case library,and the proposed ideas are further discussed and analyzed in the end of the experimental results.In addition,considering the special characteristics of the new two labelled features,this paper introduces relevant theoretical concepts to support the research of labelled features,which is usefulness of evaluation,labelled information,and nonverbal information.The experimental results show that,firstly,the similarity retrieval method of integration angle and distance is more effective,it can search for more accurate,reasonable and appropriate history experience and decision knowledge,and assist medical staff to make diagnosis decision.Secondly,considering the labelled attribute information,it has certain influence on the similar case retrieval results,and also it has some theoretical and practical significance for medical staff's diagnosis and treatment decision management and case base maintenance.