<![CDATA[Diabetes becomes an infectious disease that has a very significant increase, it's increasing not only in old age or even suffered by a young age. Based on the Data and Information Center of the Ministry of Health of the Republic of Indonesia, the prevalence of diabetes is 9% in female genitals and 9.65% in male genitals and is estimated to increase to 19.9% with the addition of population age. Currently, the Machine Learning (Depp learning) approach is widely used in carrying out the calcification of medical image data. Retinopathy Diabetes image data can be used as material to build a classification model by utilizing the Convolutional Neural Network (CNN) algorithm and Residual Network (ResNet). The test results in the model developed in this study were the comparison showing the accuracy level of CNN 68.49% while the ResNet was 81.23%, in the CNN loss test 32.57% while the ResNet was 12.59%. In the Scenario of the application of Learning Rate 0.0005 also the ResNet is better than CNN, in the CNN Learning Rate Application Scenario producing an accuracy value of 68.49% while the ResNet adopts an accuracy value of 81.84% and on the CNN loss value of 20.14 while the 10.2% Resnet has a higher level of accuracy has a higher accuracy rate Compared to CNN and the application of learning rates it also affects in building a more accurate model in the case study of Retinopathy Diabetes.]]>
Copyrights © 2023
