What Is an M-Dependent Stationary Process?

  • Thread starter Thread starter New_Galatea
  • Start date Start date
  • Tags Tags
    Process
AI Thread Summary
An m-dependent stationary process is defined mathematically such that the random variables Xs and Xt are independent whenever the absolute difference in their indices |s - t| exceeds m. This property indicates that the process exhibits a limited range of dependence, specifically constrained to the last m observations. The discussion also connects m-dependent stationary processes to ARMA(0,q) processes, suggesting that when m equals q, the relationship simplifies to a moving average model. The implication is that for indices t and s where the difference exceeds m, the variables remain independent. Understanding this relationship is crucial for analyzing the structure of time series data.
New_Galatea
Messages
5
Reaction score
0
[SOLVED] m-dependent stationary process

Hi all

Could you tell me a strict mathematical definition of "m-dependent stationary process" or maybe a link to where I could find it

Thanks In Advance
 
Physics news on Phys.org
Thank you very much
 
And may be you know what is the dependence between m-dependent stationary process (where m=q) and ARMA(0,q) process?
 
Loosely, ARMA(0,q) = AR(0) + MA(q) = MA(q) which is X_t = \epsilon_t + \sum_{i=1}^m \theta_i \epsilon_{t-i} (since q = m). See this link. You should be able to work out what this implies for m-dep. stat. process. (Hint: are Xt and Xs dependent or independent when t - s > m?)
 
Last edited:
Back
Top