Penerapan Machine Learning Untuk Prediksi Produktivitas Pertanian Berbasis Data Cuaca Di Indonesia

Abstract

The agricultural sector plays a vital role in ensuring food security and economic sustainability in Indonesia. However, agricultural productivity is highly vulnerable to weather fluctuations and climate change, which significantly affect crop yields. This study aims to develop a machine learning-based predictive model for estimating agricultural productivity using meteorological data such as rainfall, temperature, humidity, and solar radiation. Historical data from 2013 to 2023 were collected from the Indonesian Agency for Meteorology, Climatology, and Geophysics (BMKG) and the Central Bureau of Statistics (BPS). Three machine learning algorithms—Random Forest (RF), Artificial Neural Network (ANN), and Support Vector Regression (SVR)—were implemented and compared using Python. Model performance was evaluated through Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Coefficient of Determination (R²). The results show that the Random Forest model achieved the best performance, with R² = 0.912, MAE = 0.318, and RMSE = 0.445, indicating a strong predictive capability. Rainfall and temperature were identified as the most influential variables, contributing over 60% of yield variation. The findings suggest that machine learning can effectively support data-driven decision-making in Indonesia’s agricultural sector, enabling more accurate crop planning and climate adaptation strategies to enhance national food resilience.