Mixed Truncated Spline and Kernel Nonparametric Regression Model on Population Growth Rate in West Nusa Tenggara Province

Abstract

This study was conducted by considering the data pattern that differs from each independent variable to the dependent variable. If only one estimator is used to estimate the nonparametric regression curve, the resulting estimator does not match the data pattern, less precise, and tends to produce large errors. Therefore, this study aimed to model the mixed truncated spline and kernel nonparametric regression on data and predict the population growth rate in West Nusa Tenggara Province by considering the Mean Absolute Percentage Error (MAPE). Based on the study conducted, relationships generated by x 1,x 2, x 7, x 8 formed specific patterns and were indicated to follow the spline approach’s characteristics. In contrast, the pattern generated by x 3, x 4, x 5, x 6, x 9 showed the data distribution with no specific pattern and were addressed to follow the kernel approach’s characteristics. The best model was obtained with the optimal bandwidth for each variable and 3 points of optimal knots. The mixed truncated spline and kernel nonparametric regression model was suitable for modeling and predicting population growth rate data in West Nusa Tenggara Province with a relatively high proportion and a high degree of accuracy.