Accuracy/smoothness/timeliness criteria through spectral decomposition

Usage

mse(aweights, sweights, density = c("uniform", "rw"), passband = pi/6, ...)

Arguments

aweights: moving_average object or weights of the asymmetric filter (from -n to m).
sweights: moving_average object or weights of the symmetric filter (from 0 to n or -n to n).
density: hypothesis on the spectral density: "uniform" (= white woise, the default) or "rw" (= random walk).
passband: passband threshold.
...: other unused arguments.

Value

The criteria

References

Wildi, Marc and McElroy, Tucker (2019). “The trilemma between accuracy, timeliness and smoothness in real-time signal extraction”. In: International Journal of Forecasting 35.3, pp. 1072–1084.

Examples

filter <- lp_filter(horizon = 6, kernel = "Henderson", endpoints = "LC")
sweights <- filter[, "q=6"]
aweights <- filter[, "q=0"]
mse(aweights, sweights)
#>    accuracy  smoothness  timeliness    residual 
#> 0.008306348 0.449956378 0.061789932 0.299548665 
# Or to compute directly the criteria on all asymmetric filters:
mse(filter)
#>                      q=6          q=5          q=4          q=3          q=2
#> accuracy    6.024615e-31 3.003396e-04 0.0022730023 0.0036954760 0.0010393071
#> smoothness  2.495366e-32 1.387928e-03 0.0049316361 0.0056152618 0.0181979732
#> timeliness  3.783119e-31 5.977841e-05 0.0001510595 0.0001125956 0.0007184473
#> residual   -4.091914e-31 1.363420e-03 0.0051414956 0.0049053853 0.0144907081
#>                     q=1         q=0
#> accuracy   0.0009371065 0.008306348
#> smoothness 0.1382592159 0.449956378
#> timeliness 0.0105376169 0.061789932
#> residual   0.0756650916 0.299548665