Marked log-Gaussian Cox process (LGCP) to predict household per capita expenditure in Badung Regency

While the poverty rate in Badung Regency is relatively low, there is a notable discrepancy in the average per capita expenditure between the northern and southern regions of the regency. Accordingly, the government is concerned with the reduction of household poverty. Directly targeting impoverished households represents a crucial strategy for addressing the underlying causes of poverty. The data on poverty produced by Statistics Indonesia (BPS) represent a regency-level poverty rate that does not accurately reflect the condition of each household. An individual is classified as impoverished if their per capita expenditure is below the poverty line. The per capita expenditure data were obtained from national socioeconomic survey (Susenas). The coordinates of households included in the Susenas can be used to predict per capita expenditure at a higher spatial resolution through marked point process-based modelling, where the mark is per capita expenditure. This research employs the marked log-Gaussian Cox process (LGCP) methodology for the purpose of modeling household-level per capita expenditure. The results revealed a concentration of high per capita expenditure households in the Kuta Subdistricts. Factors that significantly affect per capita expenditure are the number of household members, the age of the household head, and the education level of the household head with a significance level of 5%. The estimated number of poor people in Badung Regency is 1.31%. Furthermore, the modeling process yielded a prediction of the number of poor households per subdistrict.