1
$\begingroup$

I want to know which is the procedure to remove the seasonality (anual cycle, monthly cycle, daily cycle) of a time serie. I plotted the Autocorrelogram of my time series and I extracted from there the seasonality (in my case, anual cycle), but I don't know how to remove it.

$\endgroup$
1
$\begingroup$

Usually time series is thought to be represented by following construction:

series = seasonal component + trend component + irregular component

One way to deseasonalize series is to extract trend and add irregular component back to the trend.

For example R has a function stl which can be used to form a three component matrix for time series. You can take trend and irregular component and form an deseasonalized series:

https://stat.ethz.ch/R-manual/R-devel/library/stats/html/stl.html

Actually since we observer only one time series and not these unobservable components, we need many assumptions when modeling these series.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.