The prediction of Diabetes Mellitus disease, using the classification algorithm of Voting Feature Intervals

Abstract

Diabetes Mellitus (DM) is a dangerous disease that causes the death of the fourth largest in the world. Persons with diabetes continues to increase by approximately 7 million people every year. The increase was caused by changing patterns of daily life such as a lack of activity, eating habits high in fat, carbohydrates and heredity. These habits might cause blood sugar levels rise, which further causes complications such as heart disease, kidney, liver, and gastrointestinal disorders that is dyspepsia. Dyspepsia diseases have symptoms similar to DM. There for in this study, for the diagnosis of disease were used as comparison. The application of VFI5 algorithm on the patient data can give good prediction results by generating an average accuracy of 90% is based on cross validation with a 3-fold. The result of interval training VFI5 algorithm states that DM disease and dyspepsia have similar symptoms, so for the layman it is difficult to distinguish the two diseases without laboratory tests.