Model forecast performance dataset

The model prediction performance dataset primarily involves model comparison and evaluation, model effectiveness analysis, and spatio-temporal analysis of models. Among these, model comparison and evaluation involves assessing similar prediction models using metrics such as MAPE, SMAPE, MAE, RMSE, and EVS to compare their prediction accuracy for carbon emissions. Model performance analysis involves comparing the percentage improvement in accuracy between two models, which can quantitatively assess the differences in predictive performance between models. This is also a relatively intuitive method for identifying the relative performance of models. Spatio-temporal analysis of models involves the selection of lag steps and prediction steps. The choice of lag steps directly affects the model's ability to capture historical information, while the selection of prediction steps determines the model's ability to predict future time periods.