Modification of intensive care unit data using analytical hierarchy process and fuzzy C-means model

Abstract

This paper proposes a proper methodology in data modification by using AHP (analytical hierarchy process) technique and FCM (fuzzy c-mean) model in the ICU (intensive care unit). The binary data were created from continuous data using FCM model, while the continuous data were constructed from binary data using AHP technique. The models used in this study are FCRM (fuzzy c-regression model). A case study in scale of health at the ICU ward using the AHP, FCM model and FCRM models was conducted. There are six independent variables in this study. There are four cases which are considered as the result of using AHP technique and FCM model against independent data. After comparing the four cases, it was found that case 4 appeared to be the best model, because it has the lowest MSE (mean square error) value. The original data have the MSE value of 97.33, while the data in case 4 have the MSE value of 82.75. This means that the use of AHP technique can reduce the MSE value, while the use of FCM model can not reduce the MSE value. In other words, it can be proved that the AHP technique can increase the accuracy of prediction in modeling scale of health which is associated with the mortality rate in the ICU.