GSTARX MODELLING WITH INTERVENTIONS TO PREDICT THE RISK OF CHANGES IN RICE PRODUCTIVITY IN WEST JAVA

• Main Author: Alisha Ramadina, Faza
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84204
• Institution: Institut Teknologi Bandung

Rice is one of the main sources of carbohydrates in Indonesia, which is a priority for achieving national food security. Therefore, a deep understanding of the relationship between rainfall and rice productivity becomes important to predict the risk of changes in rice productivity in the future. Using space time modeling, specifically the GSTAR model, rice productivity can be modeled with itself in other locations as well as previous times. Moreover, rice productivity is certainly influenced by other factors in the form of exogenous variables, one of which is rainfall, which will form the GSTARX model. Intervention is a model used to explain events that affect variables within a certain period of time. In this study, procedures for building GSTAR, GSTARX, and GSTARX Intervention models were introduced. The best model obtained for predicting rice productivity is the GSTAR(3;1,1,2) model. The constructed rainfall variable has not yet influenced the modeling and prediction results of rice productivity. The GSTAR(3;1,1,2) model predicts that within the next year, ten locations in West Java will still have very large rice reserves.