ANALYZING OPEN UNEMPLOYMENT RATE IN JAVA USING PENALIZED SPLINE NONPARAMETRIC REGRESSION

Abstract

In regression analysis, there are three regression curve approach methods: parametric approach, semiparametric approach, and nonparametric approach. One of the estimation methods in nonparametric regression is spline regression with parameter estimation methods, namely smoothing, truncated, and penalized. Penalized spline estimation controls the smoothness of the curve so that the curve avoids stiffness and overfitting and does not require assumptions. This study aims to analyze the open unemployment rate in Java, which has the highest open unemployment rate in Indonesia, where studies using this approach have never been conducted. The study's results resulted in an additive Mean Square Error (MSE) of 4.137 with a coefficient of determination of 44.58%, indicating that explanatory variables of 44.58% could explain the open unemployment rate. Based on the parameter significance test, the factors that significantly effect the open unemployment rate are the dependency ratio, the GDP growth rate, senior high school gross enrollment, percentage of the poor population, and population growth rate.