Penggunaan Regresi Kuantil Multivariat pada Perubahan Trombosit Pasien Demam Berdarah Dengue

Abstract

Quantile regression is an extension of the regression model of conditional quantile where the distribution is derived from the response variable expressed as a co-variate function. Quantile regression can model data that contain outliers. Patterns of platelet change in DHF patients based on body temperature and white blood cells were analyzed by quantile regression using θ = 0,25; 0,50, and 0,75. Based on the parameter estimation results, the quantile θ = 0,25 and 0,75 obtained variables that affect the platelets of DHF patients are white blood cells. Significant differences from the variables in each quantile occur because of the possibility of other factors that influence the platelets of DHF patients that are not contained in the model. The difference in the influence of factors on each quantile requires an appropriate adjustment of medical measures so that efficiency can be obtained in handling DHF patients.