TIme series decomposition using R

Question

I'm using R with the forecast package to build some time-series models. Right now, i'm dealing with multiple-seasonality data, using the tbats function.

When a plot the fitted-model, i get a plot with the time-series components. My question is, what does exactly the slope component means? (I couldn't find it in the docs).

Code:

fit <- tbats(co2, use.trend= TRUE, use.box.cox = TRUE, seasonal.periods = c(12))
plot(fit)

Thanks! enter image description here

Thanks! But i don't get why it's always positive, even when the trend seems to fall... — Fernando, Dec 05 '12 at 12:27

score 3 · Accepted Answer · answered Dec 17 '13 at 19:55

The slope m indicates the trend. You probably want to use the stl or decompose functions in R to derive the slope and season.

The general equation for decomposition is x(t)=s(t)+m(t)+e(t) where t is time, m is slope or trend, s is season and e is the error term for residuals.

Seasonality is the periodic fluctuations and trend is a generalized direction over a given time

TIme series decomposition using R

1 Answers1