Random Forest Analysis of Exogenous Variables Impacting Rice Production in the Philippines

Abstract

This research examines the relationship of rice production as the endogenous variable in a production function theory that considers key exogenous factors such as fertilizer consumption, irrigation water use, agricultural machinery, poverty rate, and agricultural land area. The study reveals the interdependencies shaping rice production in the Philippines. Applying the Cobb- Douglas function enhanced through the random forest regression algorithm establishes fertilizer consumption's focal role, focusing its essential impact on rice yields. Proper allocation of irrigation, access to agriculture machinery, poverty alleviation, and effective land use appear as significant contributors to overall production, defining 98% of the variability in rice production in random forests in both the in-sample and out-of-sample results. The findings emphasize the necessity for holistic strategies in agricultural planning, aiming for targeted interventions in fertilizer management, irrigation infrastructure, mechanized farming, poverty alleviation, and land-use.