bvhar (version 2.2.2)

rmase: Evaluate the Model Based on RMASE (Relative MASE)

Description

This function computes RMASE given prediction result versus evaluation set.

Usage

rmase(x, pred_bench, y, ...)
# S3 method for predbvhar
rmase(x, pred_bench, y, ...)
# S3 method for bvharcv
rmase(x, pred_bench, y, ...)

Value

RMASE vector corresponding to each variable.

Arguments

x

Forecasting object to use

pred_bench

The same forecasting object from benchmark model

y

Test data to be compared. should be the same format with the train data.

...

not used

Details

RMASE is the ratio of MAPE of given model and the benchmark one. Let $MAS{E}_b$ be the MAPE of the benchmark model. Then

$$RMASE=\frac{mean(MASE)}{mean(MAS{E}_b)}$$

where $MASE$ is the MASE of our model.

References

Hyndman, R. J., & Koehler, A. B. (2006). Another look at measures of forecast accuracy. International Journal of Forecasting, 22(4), 679-688.