Background: Accurate global crop yield predictions are crucial for ensuring food security, effective agricultural planning and climate adaptation. However, existing machine learning and deep learning approaches lack crop-specific feature learning, uncertainty quantification and multiscale spatial contexts, which limits their application in precision agriculture. 
 
Methods: This study presents a hybrid ensemble deep learning framework integrated with a crop-aware transformer encoder with heteroscedastic uncertainty for global multi-crop yield prediction (Maize, rice, wheat, soybean) at a 5-arcminute (~9 km) global resolution. The architecture integrates three key innovations: (1) conservative feature engineering using historical yields (4-year sequences with a 3-year temporal gap), geographic coordinates, climate zone indicators alongside temporal trends and stress indicators; (2) crop-aware multi-head attention different mechanisms with crop-specific Query/Key/Value projections enabling differential pattern learning per crop and (3) heteroscedastic output heads predicting both mean yield (&mu;) and uncertainty (&sigma;) via negative log-likelihood optimization. We implemented rigorous validation using temporal splitting (training: years &le;2013; test: years &gt;2013) and geographic blocking (5-fold GroupKFold with 0.5&deg; spatial blocks) to prevent both spatial and temporal data leakage. 
 
Result: Evaluated on the GlobalCropYield5min dataset (1982-2015) across four crops with 60,000 samples in 34 years. The proposed model achieved an overall test R2 of 0.9281, RMSE of 0.585 t/ha and MAE of 0.379 t/ha, with statistical significance confirmed by a paired t-test (p=0.011). Five-fold geographic cross-validation yielded R2&nbsp;of 0.9337&plusmn;0.0074 and rice R&sup2;=0.9462 (best), maize R&sup2;=0.9268, wheat R&sup2;=0.9201 and soybean R&sup2;=0.8298. Uncertainty quantification achieved excellent calibration (Expected calibration error = 0.024), with empirical coverage matching theoretical values (68% intervals: 69.1% coverage; 95% intervals: 94.8% coverage). Regional analysis showed consistent performance across continents (R&sup2;=0.871-0.941), with data-scarce regions showing the expected performance reduction. Ablation studies confirmed that crop-aware attention contributed +3.4% to R2, multiscale spatial features contributed 5.8% and temporal sequence features contributed +3.66%.

A Multi-crop Ensemble Framework for Yield Prediction

<div>Background: Price serves as signals of relative scarcity as well as abundance of a given agricultural product. Prices of agricultural products vary from month to month and even from day to day and this constitutes a source of risk to farmers whose livelihood depends on good pricing of their products. 
 
Method: This study was conducted to identify pattern of price dynamics and to develop model for price dynamics of beans, yellow maize and local rice using monthly price data of grains from the National Bureau of Statistics (2006-2015).&nbsp; 
 
Result: The rk of the beans followed a regular decline ranging from 0.4035 to 1 while those (rk) of yellow maize and local rice follow non regular decline with the range 0.1448 &ndash; 1(yellow maize) and 0.2945 -1 (local rice) thus indicating that the data are non-stationary. The corrected Akaike information criteria (AICC) and other information obtained for the ARIMA1,2,2 were higher than those of ARIMA2,2,2. The AICC obtained for the ARIMA1,2,2 are 958.422 (beans), 1178.169 (yellow Maize) and 1211.778 (local rice). The ARIMA2,2,2&nbsp;was thence&nbsp;favored above ARIMA1,2,2.</div>

A Multi-crop Ensemble Framework for Yield Prediction

Autoregressive Integration Moving Average (ARIMA) Model for Prices of Selected Grains in the South West Nigeria