Interpretting BIC selection
Posted: Wed Aug 29, 2012 1:50 pm
Hi all,
I have a quick question for a problem that has me stumped. I have run a recursive regression using 9 variables, including a constant. I have used all possible combinations of these variables to compute 512 regressions for every time period, between 1990 and 2010.
When I use the BIC selection criterion, for the first 15 years it works well. However for the final 5 years, it is choosing the regression model that contains the constant term, c, only. Does anyone know how to interpret this result? Is there something I can do to select a more appropriate model? I have used AIC and Rbar selection criterion too, which did not have the same problem.
Many thanks for your help.
I have a quick question for a problem that has me stumped. I have run a recursive regression using 9 variables, including a constant. I have used all possible combinations of these variables to compute 512 regressions for every time period, between 1990 and 2010.
When I use the BIC selection criterion, for the first 15 years it works well. However for the final 5 years, it is choosing the regression model that contains the constant term, c, only. Does anyone know how to interpret this result? Is there something I can do to select a more appropriate model? I have used AIC and Rbar selection criterion too, which did not have the same problem.
Many thanks for your help.