Research On The Algorithm For Prediction On Grading Diagnosis Of Heart Disease With Rough Set

As the number one killer of the health of Chinese residents,heart disease has been widely concerned by Chinese residents.In order to pursue health,people put forward more and more requirements for the medical system,which puts forward new requirements for the work efficiency of doctors.At present,doctors provide face-to-face medical services through their own clinical experience,which is a relatively slow working mode.With the development of artificial intelligence technology such as big data and machine learning,and the rise of interdisciplinary research,the application of machine learning in the field of medical diagnosis has become a research hotspot.In this study,we gradually found a way to assist doctors in diagnosis,that is,medical expert system.This paper aims to find an accurate and efficient auxiliary diagnosis method in the field of grading diagnosis of heart disease.In this paper,two optimization algorithms,genetic algorithm and heuristic algorithm based on the importance of attributes,are used to reduce the attributes of heart disease data set.After that,we combine these two optimization algorithms with rough set theory,and reduce attributes again,and compare the reduction results before the fusion of rough set theory to find an accurate and effective attribute reduction path.Then we use the attribute set with the best reduction effect as the basis of further classification experiments,and use three classification algorithms,namely support vector machine(SVM),logistic regression(LR)and naive Bayes(NB)algorithm to classify the data set,and find the most suitable classification method for the data set.In this paper,two experiments were carried out.The first one is to solve the problem of whether patients have heart disease.There are 385 pieces of data that are not ill in the data set and 460 pieces of data that are ill in the data set.The second time is to carry out the classified diagnosis experiment on the disease data.After screening and supplementing the data,the heart disease is divided into four levels,each level of heart disease has 120 data.The final experimental results show that in the heart disease diagnosis and prediction problem,the performance of genetic algorithm after the fusion of rough sets is better,but in the heart disease classification diagnosis and prediction problem,the heuristic algorithm based on attribute importance degree after the fusion of rough sets is better.In the prediction of heart disease diagnosis,the sensitivity and specificity of the genetic algorithm using the fusion rough set are 2.1%and 1.3%higher than that of the traditional genetic algorithm.In the classification prediction of heart disease,the sensitivity and specificity of the heuristic algorithm based on the importance of attributes are improved by 5.2%and 2.3%respectively after the rough set is fused.This fully shows that the fusion rough set strategy is a successful strategy in feature reduction.Finally,we use three classification algorithms to classify the best reduction results in the diagnosis prediction and classification diagnosis prediction of heart disease.The results show that in the diagnosis prediction of heart disease,support vector machine algorithm performs the best,and its classification sensitivity and specificity are 90.8%and 96.1%respectively.In the prediction of classification diagnosis of heart disease,naive Bayes algorithm has the best classification results,and its sensitivity and specificity are 89.1%and 96.2%respectively.