Pemilihan Titik Knot Optimal Model Spline Truncated Dalam Regresi Nonparametrik Multivariabel dengan GCV

Abstract

One of the most frequently studied nonparametric regression model approaches is the spline truncated. This method is applied to cases of Maternal Mortality Rate because in various cases there is an increase in maternal mortality problems so that the government is expected to be more serious in dealing with and suppressing the MMR value through the programs launched or by overcoming the factors that significantly influence the high MMR value. This study aims to examine the determination of the optimal knot point of the multivariable nonparametric spline regression model using the GCV method as the optimal knot point selection method. The criteria for selecting the best model in this study using the MSE value. The results obtained are the best model suitable for AKI 2017, namely by using the GCV method which is located in the three knot linear spline, which is 0.07722 and an MSE value of 0.018. The variables that have an effect on the model are the percentage of deliveries performed with the help of medical personnel (x1), the percentage of pregnant women who receive Fe1 tablets (x2), the percentage of pregnant women implementing the K1 (x3), and the percentage of pregnant women implementing the K4 (x4).