Page 1 of 1

Very simple question regarding times series forecast.

Posted: Thu Apr 14, 2011 5:07 pm
by scott25
Okay, so basically I'm an extreme beginner to econometrics, regressions, forecasts, and using Eviews. I'm taking a college course which requires me to find the optimal model and lag structure to make a forecast for the number of people employed during the first 6 months of 2011 given data on the monthly number of employed between 1990 and the end of 2010. Basically, I'm pretty lost on where to begin. In an earlier assignment, I had a few independent variables to use and was able to create a regression and then make the forecast off of that but in this model I only have the time series data and no other variables. I'm just not sure how to find the optimal model to use or really even how to create a forecast-able model using only this one variable.

I'm sure this is a very simple question so I apologize for wasting all of your time I just am really new to all of these ideas and this program in general and any help would be appreciated.

Edit: I'm using version 7.0 of Eviews.

Re: Very simple question regarding times series forecast.

Posted: Thu Apr 14, 2011 5:57 pm
by startz
Begin by regressing the number of people employed on it's own past values. Something like:

Code: Select all

ls y c y(-1) y(-2)