Geospatial analysis of factors affecting index of rice harvest success in West Java using INLA-based Bayesian models

Rice production during the 2012-2022 period tends to decline. Many factors lead to a decline in production, one of which is the low on potential of rice success. This issue is an obstacle to achieve food security. West Java is the third largest rice producing province in Indonesia. For this reason, the aim of this research is to establish the potential value of harvest success in West Java Province and form a spatial Bayes regression model of the potential value of harvest success of rice plants in West Java whose inference using the Integrated Nested Laplace Approximation (INLA) approach. The spatial Bayes regression model created is a linear mixed model which includes spatial random effects using the BYM2 method. The explanatory variables of the model are the number of farmer economic institutions, the number of agricultural extension workers, the number of farmer groups, the human development index, and the number of disaster events from 27 districts/cities in West Java with the potential value of harvest success as the response variable. The results of the spatial model show that the variables of the number of farmer economic institutions and the number of farmer groups have a significant influence on the potential value of harvest success. From the results of spatial mapping, it can be seen that there are neighborhood relationships that influence the value of potential harvest success where the eastern region in West Java tends to have a higher probability of harvest success.