Applying the Algorithm of Fuzzy K-Nearest Neighbor in Every Class to the Diabetes Mellitus Screening Model

Heart disease is the number one killer in world. Someone who has potential to experience heart a person with Diabetes Mellitus (DM). that detected early can reduce risk of and various other complications. This study aims analyze performance model expected be an alternative for screening DM by using machine learning method, namely Fuzzy K-Nearest Neighbour Every Class (FKNNC) algorithm. The input this was clinical data from Hospital X which consisted 7 predictor variables 1 response variable. We used confusion matrix Area Under Curve (AUC) value Receiver Operating Characteristic (ROC) curve maseure FKNNC model, help Python programming language. results obtained indicate classified as “good classification” model. seen accuracy 86% F1-score 81,8%.