Update linear_models

[jstac]

The lecture linear_models.rst needs to be updated to current style and to reflect existence of new AR1 lecture added in QuantEcon/lecture-source-py#815

[Harveyt47]

Hi @jstac now that linear_models.rst(lss) has the code to generate the plots is there anymore to this issue, or is this issue a re-write of the whole lecture into a different format. I can see some differences between the AR1 lecture on lss.

For example in defining the model in lss we define the matrix's first then the equations which govern the dynamics of the model. While in AR1 it is the reverse.

Is this the type of style you mean. Just want to make sure before I make any changes.

[jstac]

Perhaps a good start is to read through the lecture carefully, after reading the AR(1) lecture, and see how well the second flows from the first.

If not very well, then please suggest some small improvements.

At minimum, this lecture should reference the AR(1) lecture, suggest it as background reading and give some indication of how it's extended.

Also, I think this should have been cut: "Later you’ll be asked to recreate this figure."

[Harveyt47]

Hi @jstac I have looked over AR(1) and linear_models (lss).
Both lectures have a similar flow in that it moves from model set up, moving average representation, distributions and the stationary distribution.
While lss has more content outside of this it moves in this order only with the exception that at the beginning it has examples to illustrate the dynamics that govern the model.
It might be nice to point out the similarities between lss and AR(1) when referring to moving average representation and the Unconditional moments, and how lss is the same as AR(1) but with more states.
For instance in reference to moving average representation ‘Notice if n=1 and m=1, then we have the same representation in the AR(1) lecture’.