Description of Invention
This research aims to decompose the contribution of socioeconomic factors toward household consumption expenditure using a Random Forest regression approach, using log per capita expenditure as the dependent variable. Unlike traditional economic model such as Ordinary Least Square regression, which assume linear relationships, Random Forest can effectively capture nonlinear patterns. The proposed machine learning model explains approximately 85% of the variation in log per capita expenditure. SHAP analysis visually demonstrates the nonlinear relationships between selected factors within the Random Forest model. Key findings including: (1) Income, household size, and educational level are major determinants of the purchasing power of household heads. (2) The Random Forest model demonstrated a nonlinear contribution of age and household size towards log per capita expenditure, contrasting with previous studies that treated them as linear. (3) Current policy should give priority to support for the households with larger sizes and lower incomes, who tend spend a higher proportion of their earnings. The support should primarily through non-cash transfers and subsidies