Detrended time series

0

My time series exhibits a trend, and I wish to separate the trend from my data so I am looking at a stationary process. How can I do this with NumXL?

Commentaires

1 commentaire
  • There are three ways to detrend a time series:

    1. Linear detrend - Run a linear regression of your time series on t (time) and subtract that trend out. Use NxTrend function with a linear trend-type. This is method is often applied for time series with a non-stochastic/deterministic trend pattern.
    2. Difference filter- Run the difference operator (i.e. DIFF) with lag order of 1. 
    3. Other filters - Define  a new variable as a complex combination of your current and lagged time series. Most popular is called Hodrick-Prescott filter - which is part of X-12-ARIMA Function in NumXL.
    0
    Actions pour les commentaires Permalien

Vous devez vous connecter pour laisser un commentaire.

Ce n’est pas ce que vous cherchez ?

Nouvelle publication