Home » Finance Faqs » How do we remove non-stationarity of a time series?
How do we remove non-stationarity of a time series?
The 2 main types of Non-stationarity:
1.Random walk with drift
eg.,
yt = μ + yt-1 + ut
To make the series stationary we need to calculate the first difference series yt-yt-1 This is also called a difference stationary process
2.Deterministic trended time series
eg.,
yt = μ + αt + ut
To make the series stationary we need to detrend the series. This is also called a trend stationary process
It is important that the we do not detrend a trend stationary process => stationarity will not be removed. If we detrend a difference stationary process => stationarity will be removed but an MA(1) process gets added to the resulting series.
