weighted_average: Averaging forecasts using weights

In yvesmauron/univariate-time-series-forecasting: Intelligent Algorithms to be Integrated in Business Intelligence solutions.

Description

Combines forecasting methods based on the validation error. See Nowotarski, J., Raviv, E., Tr\"uck, S., and Weron, R. (2014) for more examples.

Usage

weighted_average(
  val_hat,
  val_true,
  y_hat,
  error_fun = "rmse",
  weight_fun = "inverse",
  pool_limit = 3,
  ...
)

Arguments

val_hat	forcasts on the validation set
val_true	true values of the validation set
y_hat	forecasts for the test set
weight_fun	the function to determine the weights
pool_limit	how many methods should be considered for combination
...	not used
error_measure	error measure to be used for error calculation; error measures that calculate errors per horizons will lead to a horizon specific combination, error measures over all horizons on the other hand, will lead to a simple weighted combination.

Value

combined forecasts using imrse for weighting

References

Nowotarski, J., Raviv, E., Tr\"uck, S., and Weron, R. (2014). An Empirical Comparison of Alternative Schemes for Combining Electricity Spot Price Forecasts. Energy Economics, 46, 395–412.

yvesmauron/univariate-time-series-forecasting documentation built on March 2, 2020, 12:20 a.m.