DEVELOPMENT OF MODEL FOR DIABETES MELLITUS USING DECISION TREE TECHNIQUE

Abstract

The purposes of the research were to develop a diabetes mellitus model with data mining technique using the decision tree for analysis to finding the diabetes risk factors. The risk factors for diabetes are important information that will support medical agencies to strategies planning for disease prevention to direct to group target. On conducting the research, data were collected from patients at Dan Khun Thot Hospital from Nakhon Ratchasima Province during 2007-2012, there were 4,402 cases. The data are separated into 2 sets that training set and testing set were, 90:10 percent, respectively. The modeling using j48 algorithms that a decision tree technique. On evaluating a performance of models, we use an accuracy. The results showed that the performance of model developed has accuracy 76.14% and rules of classification for decision trees have 97 rules. The risk factors that may cause diabetes were age, gender, status, address, occupational, hypertension history, history of body mass index excess, smoking, drinking and family history of diabetes.