Research And Development Of Hypertension Diagnosis System Based On Naive Bayes And ABC Algorithm

Hypertension is one of the chronic diseases that seriously affect the health of Chinese residents.Early detection and prevention of hypertension can not only save medical costs,but also relieve detriment to patients.this paper analyzes and studies the application of feature selection and classification algorithm in medical diagnostic system and designs a hypertension diagnostic system based on Naive Bayes and ABC algorithm,aiming to improve the diagnostic efficiency and accuracy of hypertension for medical staffs.This system can show its diagnostic results to the medical staffs helping them diagnose hypertension.The main work of this thesis are as follows:(1)Through referencing to a large number of literatures at home and abroad,this paper designs a research route based on the current development situation and existing problems of medical aided diagnose system.(2)The artificial bee colony algorithm is used to select the feature set of hypertension,and an improved LS-ABC feature selection algorithm is proposed to improve the poor performance of the bee local optimization ability in traditional algorithms.Through Changing the local search strategy mechanism of the searching bee,the search scope of solutions to search gets further expanded and more refined,which makes the optimization performance of the algorithm is improved.The LS-ABC algorithm is compared with standard artificial bee colony algorithm and GABC.Experimental results show that the LS-ABC algorithm could improve the quality of the feature subset,convergence speed and classification accuracy,and filter out the redundant features well.So the LS-ABC algorithm is suitable for feature selection of hypertension.(3)Combining the respective advantages of Naive Bayesian classification algorithm and C5.0 algorithm,a diagnostic model of hypertension is established.Firstly,the Naive Bayesian Probabilistic Estimation and Optimization Algorithm is used to make classifications remove all the wrong classification training sets and improve the accuracy of the decision tree when it suffers over-fitting.Then,a new decision tree is constructed using the updated training set.This decision tree performs better in dealing with the dependency of attributes.then,it is used in the test data sets to get new classification results.The new model can not only embody the accuracy of the naive Bayesian Probabilistic Estimation Optimization algorithm in classification,but also has the advantage of high speed and low complexity of decision tree classification.(4)According to the integrated framework of hypertension diagnostic system and different functions of each module,a novel hypertension diagnostic system based on naive Bayesian and improved artificial bee colony algorithm is designed.The system modules are programmed and their functions have been tested,the desired goal is achieved.