<![CDATA[Diabetes is a chronic disease that occurs when the pancreas can’t produce enough insulin or when the insulin hormone can’t be used effectively by the body. The condition of the increased blood sugar, known as hyperglycemia, is a short-term impact that often occurs in uncontrolled diabetes. Meanwhile, the long-term impact of uncontrolled diabetes can cause damage to various body systems, especially blood vessels and nerves. Early detection of diabetes in individuals who are susceptible to diabetes is the main key to control diabetes issues. Artificial intelligence can help this issue. Early diabetes detection with artificial intelligence can predict whether a person in the next 5 years has the potential to suffer from diabetes type 2 or not, based on six variables including 2-hour plasma glucose concentration in the oral glucose tolerance test, diastolic blood pressure, fold thickness triceps, body mass index, diabetes pedigree function, and age. The prediction was built by using similarity classifier based on p-probabilistic extension, trained with the Pima Indian Diabetes dataset with women as research subjects. The contribution of this research is to select representative features in the Pima Indian diabetes dataset then implement them with similarity classifier based on P-Probabilistic Extension. The aim of this study is to compare similarity classifier algorithm with K-nearest neighbor as classifier that widely used in Pima Indian diabetes dataset. The test scenario is carried out by dividing 70% of the training data and 30% of the testing data, then the accuracy for the Pima Indian diabetes data will be compared with K-nearest neighbor and the similarity classifier. Accuracy shows a success value of 75.38%, so the similarity classifier that is built can be used to predict potential diabetes with better performance than K-nearest neighbor.]]>
Copyrights © 2023
