Research And Application Of Data Mining In Chronic Diseases

At present,chronic diseases have become a factor that endangers people’s healthy life,especially chronic kidney disease.Each year,tens of thousands of people are suffering from the disease.In order to improve this situation,to alleviate the harm caused by chronic kidney disease,for chronic diseases.Predictive research has become a hot topic.In this paper,we use the modified Apriori algorithm to process the physical examination data of chronic kidney disease,excavate a rule with practical reference value to achieve the goal of predicting chronic kidney disease,and reduce the time for acquiring knowledge,which has certain practical significance.The main contents of the thesis include: 1.The research on the concept of the research topic and several kinds of more popular algorithms,the development of the corresponding data processing technology at home and abroad,and its application in the field of smart medical care are related to data mining.The medical application has a clear understanding;2.It studies the naive Bayesian classifier,decision tree and BP neural network algorithm commonly used in data mining,and analyzes their respective advantages and disadvantages.To predict the research goals of chronic kidney disease,we proposed the use of association rules Apriori algorithm for specific data mining and analysis;3.In order to solve the bottleneck in the calculation of the traditional Apriori algorithm,scan the database multiple times and generate too many candidate candidate sets,Apriori algorithm based on vertical data format is proposed respectively based on the pointer array and difference set optimization DSE algorithm and Hash table optimization based on the HE algorithm,DSE algorithm to reduce the size of the TID set by introducing a difference set to save memory while reducing the intersection solution time,HE algorithm uses Hash tables to calculate high-speed data intersection quickly and efficiently To reduce the cycle time and traversal time when calculating the intersection,compared with the traditional algorithm,these two optimization algorithms have significantly reduced the running time and improved the operating efficiency of the algorithm;4.The physical examination data of chronic kidney disease Perform preprocessing operations,convert the original data table into data tables that can be actually tapped for association rule mining through steps such as discretization and label mapping,and use multinomial logistic regression analysis algorithms to verify the results and verify the reliability of the algorithm..The valuable knowledge rules obtained through mining association rules can be used to predict the probability of chronic kidney disease to a certain extent.It can be used as a reference for doctors to diagnose diseases,and it has important significance for automated medical treatment.