Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | "Lodh Ashish" <Ashish.Lodh@Generali.com> |
To | <statalist@hsphsun2.harvard.edu> |
Subject | st: Least square regression involving time lags in the parameters |
Date | Tue, 13 Jul 2010 14:09:52 +0200 |
Dear all, I am trying to model a regression on a dependent variable using 6 independent parameters. All these variables are time series sampled at identical time intervals. I have the independence to lag the 6 parameters individually by 0 to 10 time steps in order to get a better fit. My aim is to find the best fit (with significantly low p-values for the betas) of regression by using different lags on the parameters. For example, I might get the best fit (say adjusted R-square of 94%) using the lags 0, 2, 0, 5, 5, 1 respectively on the six variables. I made a program in MATLAB using nested loops and it works fine. The only problem being large number of OLS regressions take a long time to compile. I wish to do the same in Stata. Can anybody throw some light on how it can be done in Stata? I hope there is a simpler process in Stata to do the same. Thanks Ashish * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/