AleaMid

The solution includes advanced algorithms for mid term hourly forecasting.

• The solution is based on AleaSoft exclusive forecasting model, named AleaModel, outperforming current forecasting methods.
• AleaModel results from the latest advances in Statistics and Neural Networks. It is a hybrid model that combines Neural Networks, SARIMA and Regression, taking advantage of the benefits of each individual method.


• The resulting model of this hybridation is a Neural Network model with a SARIMA structure, that overcomes traditional Neuronal Networks limitations:
  • The SARIMA structure leads to a transparent model and guarantees the selection of an optimal model.
  • The model parameters are automatically updated, they do not need manual training.
• The adaptive scheme adjusts the Neural Network parameters to the more recent performance on a on-going basis. This adaptive strategy allows the model reaction when changes in the tendency occur, avoiding the need of occasional training.

AleaModel has been proven and demonstrated: it hss been successfully used in important utilities and TSO's and improving current forecasting methods.