APPLICATION OF TRUNCATED SPLINE AND FOURIER SERIES IN MODELING THE OPEN UNEMPLOYMENT RATE IN KALIMANTAN ISLAND

Abstrak

Unemployment is a complex problem because it is one of the benchmarks for measuring the success of a region's economic development. According to the National Medium-Term Development Plan 2020-2024 and the Government Work Plan Summary 2023, Kalimantan Island is expected to have a maximum Open Unemployment Rate of 3.4%. However, provinces in Kalimantan Island still have an Open Unemployment Rate above 3.4%. Therefore, an analysis is needed to model the Open Unemployment Rate. The analytical method used is nonparametric regression, as the correlation pattern between Open Unemployment Rate and each predictor variable forms a random pattern. The estimators to be used are the truncated spline and fourier series. This study aims to compare the truncated spline and Fourier series methods in modeling the Open Unemployment Rate and determine the variables that affect Open Unemployment Rate on Kalimantan Island in 2023. Both the truncated spline and fourier series estimators have good flexibility. Based on the modelling results, it was concluded that the truncated spline method is better. The truncated spline model yields a lower GCV value (1,67545), a higher R² (0.57399), and a smaller MAE (0.821277) than the Fourier series model. Based on the analysis results, all predictor variables used, namely the Human Development Index, District Minimum Wage, population, and economic growth, significantly effect the Open Unemployment Rate with a coefficient of determination of 57.4%.