Boosting Maize Prediction: A New Approach for Farmers

Predicting maize yields is tricky, especially with elite hybrids like ZhengDan 958. Machine learning has been a game-changer, with Random Forest (RF) as a popular tool for its data handling and non-linear relationship discovery. However, new methods are needed due to the complexity of management practices affecting yields. This study explores gradient boosting, particularly Extreme Gradient Boosting (XGB) and Gradient Boosting Regressor (GBR), to predict ZhengDan 958 yields.

The team compared multiple machine learning algorithms, selected features, fine-tuned models, and visualized results. Using a dataset of over 1700 yield pairs, they found impressive accuracy. XGB and GBR predicted well, highlighting the importance of weather conditions in yield prediction. While all algorithms fared well, XGB (RMSE = 0. 37, R2 = 0. 87, MAE = 0. 26) and GBR (RMSE = 0. 39, R2 = 0. 86, MAE = 0. 27) stood out, stressing the environment's impact on crop yield. These findings open doors for further research in predictive farming.

Actions