Main Article Content
Abstract
ABSTRACT
Diabetes is a chronic disease that occurs when the pancreas does not produce enough insulin, or when the body can not effectively use the insulin that is produced. Diabetes mellitus can be divided into two types: Type 1 diabetes mellitus and diabetes mellitus type 2. This study aims to detect diabetes mellitus and may predict the development status (Metabolic Syndrome) using Learning Vector Quantization. The data needed to detect type 2 diabetes are blood sugar levels, genetics, age, physical activity, diet, smoking habits, body mass index, gender and abdominal circumference. In addition, the data also used HbA1C and cholesterol levels to detect the status of the development of type 2 diabetes mellitus (Metabolic Syndrome). The classification process is divided into two stages: stage 1 to determine the type 2 diabetes or Non diabetes mellitus, and phase 2 to predict the prognosis of type 2 diabetes into Metabolic Syndrome or Non Metabolic Syndrome (the patient is still in the category of type 2 diabetes) performed on 200 data respectively divided into 80 training data and 120 testing data. Best detection results at stage 1 that is equal to 96.67% can be obtained using learning rate (α) of 0.7, and the rate of decrement (decα) of 0.75.While the best detection results at stage 2 average accuracy rate of 92.5% using a variety of learning rate (α) and the rate of decrement (decα). Error detection in stage 2 occurs only in the Metabolic Syndrome data detected as Type 2 diabetes mellitus.
Keywords : Accuracy, Diabetes Mellitus, Learning Vector Quantization