Research On Intelligent Coronary Stent Assisted Selection Method

The traditional selection of coronary intervention stents is mostly based on the relevant knowledge and years of experience of clinicians,which is greatly influenced by subjective factors,and the effect of surgery is closely related to the medical level of doctors themselves.Moreover,the imparting of this experience knowledge is still oral,so it is impossible to transfer the knowledge of stent selection.For effective integration,management and inheritance,a quantitative method is urgently needed to guide doctors in type selection.Therefore,the research on computer-assisted coronary stent selection is of great significance in inheriting the expert knowledge in the field of interventional surgery and improving the effect of surgery.In order to solve the above problems,the paper designed the knowledge base and inference mechanism by acquiring case samples and domain expert knowledge and experience,and a hybrid inference model for coronary stent selection was constructed,which is combined with expert system,neural network and genetic algorithm.The main research contents of this paper include:Firstly,the concepts,structures and characteristics of expert system,neural network and genetic algorithm are analyzed.Then,by comparing the advantages and disadvantages of the three technologies,a hybrid inference model of expert system and BP neural network based on genetic algorithm optimization is constructed,and its overall workflow is discussed.Secondly,according to the knowledge representation method and knowledge acquisition approach of expert system,case characteristics are selected and coded by analyzing the disease information affecting the selection of coronary stents and stent characteristic parameters.Then,the acquired knowledge is represented by the production rules and the weights of the neural network.After establishing a complete knowledge base of the support selection system,it is stored in the SQL Server database for subsequent support reasoning calls.Then,the inference part of expert system is analyzed,including inference strategy selection and inference realization,and the inference process of BP network and genetic algorithm optimized BP network is discussed in detail.The predictive accuracy of each parameter of the optimized network is above 94%,and the two methods are verified by specific clinical examples.The results show that the result obtained by the hybrid inference model is in line with actual type of stents obtained by clinical application.Finally,an assistant selection system of coronary stent is designed by using Visual Studio 2010 system integrated development environment,MFC framework technology,SQL Server 2008 database management system and MATLAB neural network toolbox,which realizes intelligent reasoning of coronary stent selection.