We collect the hourly infeed data of the Dutch grids Noord and Brabant from 1/1/2007 to 12/31/2010. We use the data from 1/1/2007 to 12/31/2009 to train the models and use the data from 1/1/2010 to 12/31/2010 to test the predictive power of the models. For each grid we develop a slightly different model as the demand dynamics of each grids are not identical, for example the Dutch Carnival has a significant influence on the Brabant grid, while it has less effect on the other grid. We also take the natural logarithm of
the data to reduce the possible impact of heteroscedasticity. Below we explain the independent variables tested and used in our models.