Modelling of SDG’s Achievement in East Java Using Bi-responses Nonparametric Regression with Mixed Estimator Spline Truncated and Kernel

Abstract

The Government of Indonesia has grouped 17 objectives of SDG’s into 4 pillars, among others: social pillars, economic pillars, the pillars of environmental development and the pillars of legal development and Governance. Some strategic indicators are used to know the achievement of SDG’S especially achievement in social and economic pillars such as the Human Development Index (HDI) and Gini Ratio. There are several variables affecting then. A method used to model the effects of its variables is regression analysis. The results of the initial exploration through the scatter plot shows a pattern that changes and disconnects at a certain point for a variable IPM with a variable that is allegedly influencing it. With a scatter plot, it can be found that between the Gini ratio variables with some of its predictive variables, does not have a specific pattern. Bi-responses nonparametric regression is an analysis that appropriates with that case. Bi-responses nonparametric regression analysis is a type of nonparametric regression analysis used when the pattern between two response variables and some of the predictor variables is unknown or if there is no prior information regarding the pattern of them. Most of researchers use bi-responses nonparametric regression approached by spline truncated, kernel and etc. Most of the research, use kernel and spline truncated because of their advantages such as: have a good capability to develop model from unknown pattern data and have a good visual representation. The first aim of this paper is obtaining the estimator of mixed estimator spline truncated and kernel in bi-responses nonparametric regression with weighted least square. The second aim is gaining a best model to explain the relationship indicators of SDG’s achievement with some variables affecting them. To obtain the best model, it is necessary to select the optimal of knots point and bandwidth parameters using a smallest Generalized Cross Validation (GCV).