Hi, I'm trying to create a random walk model with -arima-. What I want is that L1 (phi) is equal to 1, because then you have a random walk.
I have the following code:
Where the first 750 observations are the in-sample period and the rest is the out-of-sample period. With this code, I get an L1 close to 1, but not exactly 1.
Does someone know how to get a L1 of exactly 1?
I have the following code:
Code:
arima variable in 1/750, noconstant arima (1,0,0) predict variable_fcst_RW in 751/1975, xb
Does someone know how to get a L1 of exactly 1?