Mathematical modeling of gross regional domestic product growth rate of North Sumatra Province by business field using local polynomial regression

Abstract

Nonparametric regression modelling for data applies the relationship between predictor and response variables without considering any particular trend. The main principle of nonparametric regression is estimating an unknown smooth function. Local polynomial regression is one of several methods for estimating smooth functions used in nonparametric regression models. There are two parameters in the local polynomial regression model the smoothing parameter and the polynomial degree parameter. Generalized Cross Validation is a classical method used to determine optimal smoothing parameter in nonparametric regression. The smoothing parameter that gives the minimum value of Generalized Cross-Validation is the optimal parameter. We apply the optimal smoothing parameter for local polynomial regression modelling using data on the regional domestic product growth rate of the business field in North Sumatra Province. The simulation results show a polynomial of degree two is better than degree one. Investigation of the effects of the Covid-19 pandemic shows that the predicted growth rate is far from the expected growth.