Automated Forecast Combination
Computes the fit for all the available forecast combination methods on the provided dataset with respect to the loss criterion. Returns the best fit method.
auto_combine(x, criterion = "RMSE", param_list = NULL)
x: An object of class 'foreccomb'. Contains training set (actual values + matrix of model forecasts) and optionally a test set.criterion: Specifies loss criterion. Set criterion to either 'RMSE' (default), 'MAE', or 'MAPE'.param_list: Can contain additional parameters for the different combination methods (see example below).Returns an object of class foreccomb_res that represents the results for the best-fit forecast combination method: - Method: Returns the best-fit forecast combination method.
Models: Returns the individual input models that were used for the forecast combinations.
Weights: Returns the combination weights obtained by applying the best-fit combination method to the training set.
Fitted: Returns the fitted values of the combination method for the training set.
Accuracy_Train: Returns range of summary measures of the forecast accuracy for the training set.
Forecasts_Test: Returns forecasts produced by the combination method for the test set. Only returned if input included a forecast matrix for the test set.
Accuracy_Test: Returns range of summary measures of the forecast accuracy for the test set. Only returned if input included a forecast matrix and a vector of actual values for the test set.
Input_Data: Returns the data forwarded to the method.
The function auto_combine allows to quickly apply all the different forecast combination methods onto the provided time series data and selects the method with the best fit.
The user can choose from 3 different loss criteria for the best-fit evaluation: root mean square error (criterion='RMSE'), mean absolute error (criterion='MAE'), and mean absolute percentage error (criterion='MAPE').
In case the user does not want to optimize over the parameters of some of the combination methods, auto_combine allows to specify the parameter values for these methods explicitly (see Examples).
The best-fit results are stored in an object of class 'foreccomb_res', for which separate plot and summary functions are provided.
obs <- rnorm(100) preds <- matrix(rnorm(1000, 1), 100, 10) train_o<-obs[1:80] train_p<-preds[1:80,] test_o<-obs[81:100] test_p<-preds[81:100,] data<-foreccomb(train_o, train_p, test_o, test_p) # Evaluating all the forecast combination methods and returning the best. # If necessary, it uses the built-in automated parameter optimisation methods # for the different methods. best_combination<-auto_combine(data, criterion = "MAPE") # Same as above, but now we restrict the parameter ntop_pred for the method comb_EIG3 to be 3. param_list<-list() param_list$comb_EIG3$ntop_pred<-3 best_combination_restricted<-auto_combine(data, criterion = "MAPE", param_list = param_list)
Christoph E. Weiss and Gernot R. Roetzer
foreccomb, plot.foreccomb_res, summary.foreccomb_res, accuracy