Application of Generalized Space Time Autoregressive Model on Farmer Exchange Rate Data in Three Provinces of The Sumatera Island

• Main Authors: Fadhilatul Nida Aryani, Sri Sulistijowati Handajani, Etik Zukhronah
• Format: Article
• Language: English
• Published: Fakultas MIPA Universitas Jember 2020-07-01
• Series: Jurnal Ilmu Dasar
• Online Access: https://jurnal.unej.ac.id/index.php/JID/article/view/17226

The agricultural sector has a big role in the development of the Gross Regional Domestic Product (GDP). Therefore the agricultural sector is very important. Besides the agricultural sector, the farmer's welfare also needs to be considered because the agricultural sector will be good if the welfare of farmers is good also. In measuring the level of farmers' welfare, the method used is the farmer's exchange rate. The farmer's exchange rate has a location relationship and a previous time relationship. The Generalized Space-Time Autoregressive (GSTAR) model is a good method of forecasting data that contains time series and location relationships by assuming that the data has heterogeneous characteristics. The purpose of this study is to model the farmer exchange rate data with GSTAR using normalization of cross-correlations weighting and inverse distance in three provinces namely West Sumatra, Bengkulu and Jambi Provinces. Based on data analysis, the best GSTAR model obtained by using the best weighting with the model is GSTAR (11) − I(1) using normalization of cross-correlations because the assumption of normal white noise and multivariate are fulfilled with an RMSE value of 1.097775. The best GSTAR model explains that the exchange rate of West Sumatra farmers is only the previous time, Bengkulu farmers' exchange rate is the previous time and is the exchange rates of farmers of West Sumatra and Jambi, whereas for the exchange rate of farmers of Jambi is the exchange rates of farmers of Bengkulu and West Sumatra and influenced by previous times. Keywords: GSTAR, RMSE, farmers exchange rate, normalization of cross-correlations, inverse distance.