vangspiliot / AutoTheta

Generalizing the Theta method for automatic forecasting
3 stars 2 forks source link

AutoTheta

Generalizing the Theta method for automatic forecasting

The Theta method, as originally proposed by Assimakopoulos and Nikolopoulos (2000), is generalized for automatic forecasting by introducing three extensions on its framework to boost its performance. This includes (i) considering both linear and non-linear trends, (ii) allowing to adjust the slope of such trends, and (iii) introducing a multiplicative expression of the underlying forecasting model along with the existing, additive one.

More specifically, AutoTheta is capable of the following:

Example of use

AutoTheta is developed in R and can be used for forecasting a ts() object as follows:

library(Mcomp)

tsi=1100

time_series <- subset(M3, 12)[[tsi]]

frc <- AutoTheta(time_series$x, fh=18, positive=TRUE)

plot(time_series)

lines(frc$mean, col="blue")

lines(frc$piu, col="blue")

lines(frc$pil, col="blue")

References