Predicting Diabetes Disease Using Data Mining Classification Algorithms and Comparison of Algorithm Performances

Abstract

Data is always produced everywhere in the universe. In our globalizing world, where the internet is rapidly spreading, countless data set examples can be given. We produce data when we shop at a bookstore, make a transfer from our bank account, travel with an airline, or get tested in a hospital. Test results of hospital patients, production reports of a manufacturing factory, and exam results of students are examples of data sets. This data pile grows daily, and the data has no meaning. Data gains meaning by being processed by various tools and transformed into information. Today, information has become the most incredible power. In this study, a prediction study of diabetic patients was made with ten different algorithms from classification methods, one of the data mining methods. At the end of the prediction study, the algorithms' performances were compared. The unique aspect of this study is that the performance of the algorithms is compared by making predictions with ten different classification methods instead of just one or a few classification methods.