**3. Results**

A total of 281 diabetes patients were evaluated; 121 (43.06%) were male and 160 (56.93%) were female. Among the 281 records, 256 (91.10%) were not dependent on insulin (Type 1), 14 (4.98%) were Gestational, and 11 (3.91%) were insulin dependent (Type 2). Initially, the dataset was divided into a 20:80 ratio for conducting training and testing. After training the machine, a 10-fold cross-validation technique was implemented on an experimental platform of Weka for better assessment of the classification. The dataset was divided into 10 samples. Each sample was utilized as validation data from the retention process, while the remaining nine samples served as the training data. This process was performed 10 times. The advantage of this process is the reduction in the error ratio and bias correlation by random sampling.
