Model Regresi Kuantil Spline Orde Dua Dalam Menganalisis Perubahan Trombosit Pasien Demam Berdarah

Abstract

Quantile regression can be used to analyze data containing outliers including DHF data. The spline is able to identify several patterns of change in the regression model, so this study uses a second-order quantile spline regression model in analyzing DHF data that occurred in Makassar City. In this article, the authors analyze the pattern of changes that occur in platelets based on changes in the hematocrit content of DHF patients. The selected quantiles are quartiles 0.25; 0.50; and 0.75 with 3-knot points. Based on the results of the analysis, the minimum GCV value obtained at the use of knot points is 30.30; 44.80; 47.10 for the 0.25 quartile; 0.50; and 0.75. This shows that in each quartile, there are four patterns of quadratic changes that occur in the platelet count of DHF patients. The parabolic curve formed in each pattern segmentation shows that there are times when platelets are increasing and there are times when platelets are decreasing. However, the average platelets decreased drastically, especially when the hematocrit reached 47.10%.

Keywords

• hematocrit
• quantile regression
• second order spline
• platelets