IMPLEMENTASI ALGORITMA CART DALAM KLASIFIKASI PENYAKIT DIABETES

Abstract

Diabetes is a metabolic disease caused by a lack of insulin production in the pancreas, this results in an imbalance of sugar in the blood so that the concentration of blood sugar levels increases. Patients with diabetes from year to year are increasing. Estimates from the International Diabetes Federation (IDF), there are 382 million people suffering from diabetes in 2012. It is estimated that by 2035 the number will increase to 592 million people. Recording of this disease needs to be done so that prevention can be done. One of the records that can be done is by utilizing data mining classification techniques. This study implements the CART (Classification And Regression Trees) algorithm in the classification of diabetes. The highest accuracy results were obtained when classification using the CART algorithm without pruning and prepruning was 100%. Meanwhile, pruning and prepruning produce an accuracy of 96.15%.
Keywords : data mining, classification, diabetes, CART.