Model Multivariate Adaptive Regression Spline (MARS) Pada Persamaan Regresi Nonparametrik

Abstract

Economic growth is one of the benchmarks for the success of development or increasing welfare in the government of a region in the economic sector as measured by the Gross Regional Domestic Product (GRDP). It has time series data that often fluctuates so that the appropriate method is nonparametric regression. This study also aims to determine the most influential factors for economic growth in North Sumatra in 2019-2021 using the MARS model, using secondary data published by BPS for 2019-2021. The MARS model is obtained by obtaining a combination of BF, MI, and MO values that have a minimum Generalized Cross Validation (GCV) value. The results of this study indicate that the best MARS model is a combination of BF=28, MI=1, and MO=1 with a GCV value of 8.42E+06. Therefore there are five of the seven variables that have a significant effect on economic growth in North Sumatra, namely population (X_7 ) with an interest rate of 100%, domestic investment (X_5) of 76.86%, local revenue (X_1) of 31.14%, allocated funds special (X_3) of 28.89%, general allocation funds (X_2) of 23.14%.