Efficient Crop Yield Forecasting Using LightGBM with Soil and Climatic Indicators

Abstract

Accurate crop yield prediction is a critical component of agricultural planning, food security, and sustainable farming practices. However, crop yield is influenced by complex and nonlinear interactions among environmental conditions and agricultural management factors, which are often inadequately captured by traditional statistical models. Motivated by these limitations, this study proposes a machine learning–based approach for crop yield forecasting using an ensemble regression model that integrates climatic variables and agricultural management indicators. The objective of this research is to develop an efficient and reliable prediction model capable of learning complex patterns from heterogeneous agricultural data. The proposed methodology involves data preprocessing, feature selection, model training using an ensemble learning framework, and comprehensive performance evaluation. Experimental results demonstrate that the proposed model achieves strong predictive performance, as indicated by favorable values of R-squared (R²), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE). A visual comparison between actual and predicted yield values further confirms the model’s ability to capture overall yield trends and generalize well to unseen data. In addition, feature importance analysis provides insights into the relative influence of climatic and management-related factors on crop yield. The main contribution of this study lies in the integration of multiple influential factors within a single ensemble-based prediction framework, offering a more comprehensive and accurate approach to crop yield forecasting. Future work will focus on incorporating additional agronomic variables, extending the model to multi-crop and multi-region datasets, and exploring advanced hybrid and explainable machine learning techniques to further enhance prediction accuracy and interpretability.