Re: Linear regression or time-series forecasting? What's more appropriate?

["Reef Fish"]

Nick wrote:

< very lengthy question snipped>
>
> From memory (30 years ago) and searching on Google, much of time-series
> forecasting uses linear regression.

This is a completely inappropriate characterization as well as
description
of how time-series data are used in forcasting!

Some kinds of time series models CAN be approximated by the use of
lag and differences as independent variables in a regression.   My
co-author Harry V. Roberts have written books using that approximate
method, but that is hardly the correct nor the usual way of using
time series methods in forecasting.

-- Reef Fish Bob.