Comparison between Multiple Linear Regression and Fuzzy C-Regression Models towards Scale of Health in ICU


Rusiman M. S., Adnan R., NASİBOĞLU E., Jacob K.

MATEMATIKA, vol.27, no.2, pp.183-198, 2011 (ESCI) identifier

  • Publication Type: Article / Article
  • Volume: 27 Issue: 2
  • Publication Date: 2011
  • Journal Name: MATEMATIKA
  • Journal Indexes: Emerging Sources Citation Index (ESCI)
  • Page Numbers: pp.183-198
  • Keywords: Multiple linear regression (MLR) model, fuzzy c-regression models (FCRM), mean square error (MSE)
  • Dokuz Eylül University Affiliated: Yes

Abstract

The multiple linear regression (MLR) model is well-known in analyzing linear model. Whereas, the new technique in clustering data, the fuzzy c-regression models (FCRM) are being widely used in analyzing the nonlinear model. The FCRM models are tested on simulated data and the FCRM models can approximate the given nonlinear system with a higher accuracy. A case study in scale of health at intensive care unit (ICU) using the two methods of modelling as mentioned above was carried out. The comparison between the MLR and FCRM models were done to find the better model by using the mean square error (MSE). After comparing the two models, it was found that the FCRM models appeared to be the better model, having a lower MSE. The MSE for MLR model is 498.29 whereas the MSE for FCRM models is 97.366.