Short-term load cross-forecasting using pattern-based neural models

In this article we present the idea of short-term load cross-forecasting. This approach combines forecasts generated by two models which learn on input data defined in different ways: as daily and weekly patterns. Pattern definitions described in this work simplify the forecasting problem by filtering out the trend and seasonal variations. The nonstationarity in mean and variance is also eliminated. Simplified relationships between predictors and output variables are modeled locally using one-neuron models. A simulation study on the sample of real data showed better performance of cross-forecasting than individual neural models.