Authors: Ambuj Kumar Misra
Abstract: Global food security remains a critical challenge in the 21st century, with crop yield prediction serving as a fundamental tool for agricultural planning and resource management. This research presents a comprehensive analysis of machine learning-based approaches for predicting crop yield using secondary data from the Food and Agriculture Organization (FAO) and the World Bank. This study evaluates multiple machine learning algorithms including Random Forest, Gradient Boosting, Support Vector Machines, Neural Networks, XGBoost, and Long Short-Term Memory (LSTM) networks on a dataset comprising 15 years of agricultural data across 45 countries. Results demonstrate that XGBoost achieves the highest predictive accuracy at 91.3%, with a Root Mean Square Error (RMSE) of 0.487 tons per hectare. The integration of climate variables, soil characteristics, and socioeconomic indicators significantly improves model performance.
DOI: http://doi.org/10.5281/zenodo.21452546
