Compute mean absolute scaled error (MASE)

Computes the Mean Absolute Scaled Error (MASE) between a reference (naive) prediction and one or more alternative predictions. MASE is scale-independent and robust to outliers, making it useful for comparing forecast accuracy across different indices or time series.

Usage

mase(ref, preds, ...)

# S4 method for class 'FLQuant,FLQuants'
mase(ref, preds, order = c("inverse", "ahead"))

# S4 method for class 'FLIndices,list'
mase(ref, preds, order = "inverse", wt = "missing")

Arguments

ref: Reference or naive prediction, an FLQuant or FLIndices time series.
preds: Predictions to compare to the reference; an FLQuants or a list of FLIndices.
...: Additional arguments passed to methods.
order: Character; whether predictions are in "inverse" (default, most recent first) or "ahead" order.
wt: Mean weights-at-age to use with indices.

Value

A numeric scalar (or named numeric vector for the FLIndices, list method) giving the MASE value(s).

Generic function

mase(ref, preds, ...)

References

Franses, P.H. (2016). A note on the Mean Absolute Scaled Error. International Journal of Forecasting, 32(1):20–22. doi:10.1016/j.ijforecast.2015.03.008 .

Author

The FLR Team