Mirrors morie.fn.joholt. With damped = TRUE the trend
contribution is \(\sum_{i\le h}\phi^i\) rather than h, so long
horizons flatten instead of extrapolating a line forever.
Usage
morie_holt_linear(
y,
alpha = NULL,
beta = NULL,
horizon = 1L,
damped = FALSE,
phi = 0.98
)
Arguments
- y
numeric series, at least 5 observations.
- alpha, beta
smoothing parameters in (0, 1); estimated by
one-step SSE when NULL.
- horizon
forecast horizon.
- damped
logical; damp the trend.
- phi
damping parameter in (0, 1].
Value
list: forecast, level, trend, fitted, residuals, sse,
alpha, beta, damped, n.
References
Hyndman, R. J. & Athanasopoulos, G. (2021).
Forecasting: Principles and Practice (3rd ed.). OTexts. Sec. 8.2.
Examples
morie_holt_linear(1:30 * 2, horizon = 3)$forecast
#> [1] 62 64 66
