Machine Learning-Based Crop Yield Prediction for Sustainable Agriculture in Eastern Ethiopia

Abstract This study focuses on developing a machine learning-based crop yield prediction model suitable for agricultural conditions in Eastern Ethiopia, presents research that has taken cognizance of the fact that agriculture occupies the leading role in the economy through providing a source of livelihood for most Ethiopians and being one of the main contributors to the nation's GDP. Thus, addressing challenges such as low productivity and food insecurity induced by climate change and environmental factors becomes paramount. Agriculture in Ethiopia is very sensitive to changes in weather conditions, soil degradation, and other ecological factors; hence, exact yield forecasting is quite indispensable for proper planning and resource management of the sector. It integrates various data sets that include but are not limited to local agri-data, historical yield records, and satellite-derived environmental information. Advanced machine learning algorithms have been considered in the model, such as Random Forest, Gradient Boosting, K Neighbors, and Decision Tree regressors for the prediction of yields precisely. The methodology involved rigorous data preprocessing, feature selection, and model training to give robustness and reliability to predictions. Results clearly indicated that the Random Forest Regressor model outperformed all other alternative models and showed its prospect in enhancing agricultural productivity by informing decisions to be taken by the farmers and other stakeholders. This research underlines technology utilization as a means of optimizing crop management practices, one of the crucial pre-conditions for improved food security and sustainable agricultural development in the region. The results of this study do not end at the immediate farming outcomes, but also remind that technology-driven interventions may potentially address broader socio-economic problems in Eastern Ethiopia.