Prediction of Paddy Production in Indonesia Using Semiparametric Time Series Regression Least Square Spline Estimator

Support for one of the points of Sustainable Development Goals (SDGs), namely Zero Hunger, is by supporting sustainable agricultural empowerment. Indonesia is one of the countries with the fourth largest rice consumption according to the United States Department of Agriculture. 90% of Indonesians consume rice as a staple food. In this study, we model paddy production in Indonesia using a semiparametric time series regression approach based on least square spline estimator (LSSE). Where spline is used to overcome data that tends to fluctuate in monthly paddy production data. Monthly data on paddy production in Indonesia over a certain period of time is used to build a model. The use of a semiparametric regression approach by combining parametric components and nonparametric components for analyzing factors that affect paddy production. In this study, the parametric component is paddy production in the previous period lag-1 and the nonparametric components are the potential area of crop failure and the generative area. For predicting paddy production in Indonesia using Semiparametric Time Series Regression Model (STSRM) approach based on LSSE, we determine the order and optimal knot points based on the smallest Generalized Cross Validation (GCV) value. The results of the study show that the Mean Absolute Percentage Error (MAPE) value of 18.05% is less than 20%. It means that prediction of paddy production in Indonesia using STSRM based on LSSE is a good prediction