I am a little confused how AR models are different from random walk problems?
In random walk, values depend on previous value +/- error term. The same seems to be true for AR models?
What am I missing here?
The coefficient is of the independent term is 1 in case of random walk models and other than 1 in other AR models. Random walk model is just one type of AR models.