<![CDATA[Indonesia, as an agrarian nation, relies heavily on agriculture for rural livelihoods. The Farmer Terms of Trade (FTT) is a key indicator of farmer welfare. However, agriculture is often seen as ineffective in boosting income and reducing poverty. Despite this, the sector remains crucial for national development, especially the crops sub-sector, which sustains the country's food supply. The Generalized Space Time Autoregressive (GSTAR) model is employed to explore data relationships across proximate locations, focusing on geographical or observational locational factors. This analysis incorporates three spatial weights in the GSTAR model: (1) queen contiguity- weights, (2) uniform location weights, and (3) inverse distance spatial weights. Our findings indicate that the GSTAR model (11)I(1) with uniform spatial weight emerges as the optimal model. This model not only satisfies the white noise and normality assumptions but also demonstrates superior performance metrics, including a Mean Squared Error (MSE) of 2.34, Root Mean Squared Error (RMSE) of 1.53, and Mean Absolute Percentage Error (MAPE) of 1.10%. These figures notably surpass those obtained with the GSTAR models employing queen contiguity-based weights and inverse distance spatial weights, thereby highlighting its efficacy in capturing the dynamics within the crops sub-sector.]]>
Copyrights © 2025
