<![CDATA[Diabetes is a chronic disease characterized by high blood sugar. As of 2011, there were 7.29 million people suffering from diabetes, and in 2021, there were 19.47 million people who have diabetes. The percentage increase in people with diabetes from 2011-2021 has a percentage increase of 267%. Very rapid growth and one of the causes of death worldwide is a problem that needs to be solved. Reduce the number of people with diabetes, there are various ways, but they are not optimal. So it is necessary to research to develop a system that can detect diabetes early so that treatment or prevention can run well. One of the techniques that can be used to detect diabetes early is prediction. The K-Nearest Neighbor (K-NN) algorithm is an algorithm designed to classify data based on previously classified learning data however this algorithm has a weakness in processing data that has high dimensions and is non-linearly separable, so adding a kernel function is a good choice for input data clustering. From the results of this study, the value of k and the kernel function with the highest accuracy value is k = 50. The kernel function Linear and Polynomial degree 1 and the performance of the Kernel K-Nearest Neighbor algorithm are better than the K-Nearest Neighbor algorithm with a difference in the accuration value of 0.14.]]>
Copyrights © 2022
