<![CDATA[Poverty in Indonesia, especially in Java, remains a major challenge despite the island being the economic and political centre of the country. The government has made many efforts but has not been effective in overcoming poverty. The hierarchical structure of poverty data may cause higher-level clusters to be random effect. One approach that can be used to represent the relationship between the poverty rate in each regency/city in Java and the factors that influence it with the province as a random effect is a linear mixed model (LMM). The number of factors that can affect poverty results in multicollinearity. The application of LASSO is used in this study to overcome multicollinearity, select, and generate variables that are significant to poverty in Java. The data used in this study consists of 85 regencies and 34 cities in Java Island involving 20 independent variables. The results show that the factors that influence the poverty rate are average years of schooling, non-food expenditure, number of households with housing assets owned, percentage of households with a dirt floor, and percentage of households with PLN lighting. The LMM-LASSO is a linear model augmented with a LASSO penalty function to address multicollinearity and incorporates random effects into the model. This approach is suitable for modeling the poverty rate, as indicated by its smaller AIC and BIC values compared to the conventional linear mixed model. In addition, based on the ICC value, the province as a random effect contributes significantly to the variability of the data at the district/city observation level in Java Island.]]>
Copyrights © 2025
