DETERMINING THE BEST DEFUZZIFICATION METHOD FOR THE MAMDANI FUZZY INFERENCE SYSTEM IN THE DIAGNOSIS OF PRE-ECLAMPSIA IN PREGNANT WOMEN

Abstract

Pre-eclampsia is the second of the three major causes of death in pregnant women after bleeding followed by infection. Pre-eclampsia is a disorder of unknown etiology specifically in pregnant women. To prevent pre-eclampsia from becoming increasingly severely diagnosed systems that can be used for pre-eclampsia premises syconome. One method that can be used to determine the pre-eclampsia diagnosis is the Fuzzy Inference System (FIS) Mamdani this method is based on the concept of fuzzy logic. The process of determining the final decision by this method has several stages, the application of implications, rules, and defuzzification composition. For defuzzification stages, there are four methods that can be used the method Centroid, Bisector,  Mean of Maximum (MOM), Smallest of Maximum (SOM), and Largest of Maximum (LOM). This study aims to determine the diagnosis of pre-eclampsia (pregnancy poisoning) in pregnant women based on FIS Mamdani by previously determining the best FIS Mamdani defuzzification method. In determining the best defuzzification method, the measures Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE), Mean Square Error (MSE), and Sum Square Error (SSE) are used. Based on the results of the prediction error comparison, the best defuzzification method to diagnose the pre-eclampsia status in Atambua Hospital is a bisector method with an accuracy of 95,48%.