Etstimate AR(1) coefficient in Matlab
20 views (last 30 days)
Show older comments
Hello, I have a time series data of GDP of US during 1980-2017 yearly.
I assume that GDP follows an AR(1) process, like: y(t) = rho*y(t-1) + e(t)
Now I want to estimate the coefficient rho. I was searching and found 2 ways:
1- use the command ar: ar(y,1) (as instructed here: https://www.mathworks.com/help/ident/ref/ar.html)
2- use the Arima estimation: md1=arima(1,0,0) Estmd1 = estimate(mdl,y) (as instructed here: https://www.mathworks.com/help/econ/arima.estimate.html) I tried both way and got different result on the estimation of coefficient rho. Could you tell me which way is correct?
Thanks!
1 Comment
mirewuti muhetaer
on 18 Nov 2019
As far as i know, ar(y,1) uses least sequared estimation (OLS) approach while estimate() funciton uses MLE approach. So the result is definitely different.
Answers (0)
See Also
Categories
Find more on Smoothing in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!