Predicting Interprovincial Rice Food Security in Indonesia as a Pillar of National Defense Using the Random Forest Regressor Algorithm

Bagus Hendra Saputra (Universtias Pertahanan Republik Indonesia, Bogor, Indonesia)
Ahmad Eryan Firdaus (Universitas Pertahanan Republik Indonesia, Bogor, Indonesia)

Journal of Defense Technology and Engineering
Vol. 1 No. 1 (2025): July, Journal of Defense Technology and Engineering

This study investigates interprovincial rice food security in Indonesia as a strategic pillar of national defense. Using a quantitative predictive approach, the Random Forest Regressor algorithm was applied to multidimensional data from all provinces, incorporating variables such as rice expenditure per capita, rice price, production, population, consumption, and harvested area. The results show significant disparities between provinces: surplus regions such as East Java, Lampung, and South Sulawesi contrast sharply with deficit areas like Jakarta, Papua, and Bangka Belitung. Feature importance analysis reveals that supply-side factors, particularly harvested area (50.5%) and production (33.2%), are the most decisive, while demand-side factors have weaker influence. Model evaluation achieved an R² of 0.8239, confirming strong predictive reliability. These findings underscore that rice food security is not only an economic and social issue but also a critical aspect of non-military defense. Strengthening predictive systems and interprovincial distribution networks is essential to ensure resilience against disruptions from disasters, conflicts, or geopolitical instability. The study highlights the practical value of machine learning models in guiding evidence-based policy to secure national food sovereignty.