Bootstrap Aggregating Multivariate Adaptive Regression Splines (Bagging MARS) dan Penerapannya pada Pemodelan Produk Domestik Regional Bruto (PDRB) di Provinsi Sumatera Barat

Abstract

Increased economic growth can help a region's economy grow and demonstrate that the government is capable of improving the welfare of its citizens. The rate of economic growth may be measured by gross regional domestic product (GRDP). This look at turned into performedto decide the factors that maximum effect GDRBinside the province of West Sumatera from 2015 to 2019 using Bootstrap Aggregating Multivariate Adaptive Regression Splines (Bagging MARS). The best model with the lowest GCV value is 7,36868 with BF=8, MI=3 and MO=0 as a combination. Then Bagging was carried out on the initial dataset with 50 Bootstrap replications to obtain the smallest GCV of 5,256292. Based on this, the smallest GCV value obtained from Bagging MARS is smaller than the MARS method. Meaning that the Bagging method can lessen the GCV value and increase accuracy. So that the factors that maximum influence GRDP in the province of West Sumatera are Regional Original Income.