Table 7

Model performance across different evaluation scopes

ModelMAEMSER2
Linear Regression3687.6475,236,811.980.465
Random Forest3474.7370,883,256.960.497
ARIMA26.341296.29−0.259
LSTM (Maize)15.36364.620.640
LSTM (Rice)11,893.29153,813,836.48−0.039

Note(s): Linear Regression and Random Forest were evaluated on the full multi-commodity dataset, whereas ARIMA and LSTM models were evaluated on selected commodity-level series. Accordingly, metric values should be interpreted within their respective evaluation contexts and are not directly comparable across all models

or Create an Account

Close subscription notice
Close access options