What is the difference between the order of an AR model and the lag? can someone explain plz?
In practical terms, they are the same, however, each one is used in different scenarios. AR means the constitution of the model, for example AR(2) is x(t) = b0 + b1*x(t-1) + b2*x(t-2) + e(t) Lag means no necessarily a model per se, it is used to describe the lenght of an analysis. For example, in error analysis, you may want to determine if correlation of errors exist, so you can check if even until the lag number 36 the errors present correlation.
I got the AR thing. But i’m still confused with “lag”. Is it the time period ?
Yes, lag means “how much time in the past you looking at”. For example, I want to know if error today has significant correlation with 3 past-period error, so I use 3rd lag of error. In math terms I calculating corr[e(t),e(t-3)].