PEMODELAN TINGKAT PENGANGGURAN TERBUKA DI SUMATERA UTARA MENGGUNAKAN REGRESI NONPARAMETRIK SPLINE TURNCATED

Abstract

This study proposed to model the unemployment rate and analyze the factors that allegedly affect the unemployment rate in North Sumatra Province. The factors that are estimated to influence the unemployment rate in North Sumatra are the economic growth rate, labor force participation rate, population density, and the percentage of poor people. The scatter plot between the open unemployment rate and the factors suspected of influencing the unemployment rate shows a tendency not to form a certain pattern, so that it can be modeled with spline truncated nonparametric regression. Truncated spline nonparametric regression can overcome sharp changes in data patterns up or down with the help of knot points, and the resulting curve is relatively smooth. In this study, modeling was carried out with a one-knot point, two-knot points, three-knot points, and a combination of knot points. The selection of optimal knot points as the best model is done by selecting the minimum Generalized Cross Validation (GCV) value. The data used is sourced from the official website of Statistics Indonesia, with observation units of 33 districts/cities in North Sumatra. The results of this study show that the best model with the minimum GCV value of 1.5237 is at the combination of knot points (1,1,1,2) with an  value of 90.78%. Based on the partial test, the economic growth rate, labor force participation rate, population density, and the percentage of poor people affect the unemployment rate.

Keywords: Generalized Cross Validation, truncated Spline; unemployment rate;