Modified Marked Log-Gaussian Cox Process for High-Resolution Prediction of Per Capita Expenditure in Badung Regency

Over the past five years, there has been no appreciable reduction in the prevalence of poverty in Badung Regency. In order to address the fundamental causes of poverty, it was necessary to obtain more detailed data. Meanwhile, Statistics Indonesia (BPS) as a national statistics agency in Indonesia has only released aggregate poverty data at the regency level. The estimation of poverty at the household level had been conducted using point process-based modeling, although this did not take into consideration uninhabited areas. This study modified the method by excluding uninhabited locations, and the resulting data were then compared with the application of the previous method. The method in this study showed results that were closer to population data at both the district and subdistrict levels. This research could be used by local governments to estimate the location of poor households as well as the number of poor households at a lower level.