Cochrane-Orcutt/AR(1) help
Posted: Wed Jul 17, 2013 3:43 am
Hi.
I am currently performing a time-series regression analysis, and need to adjust for serial correlation. I have added "ar(1)" to the end of my equation, and obtained a DW(durbin watson) value that is acceptable. However my R-squared jumped from 0.65 to 0.9-something. Can I use this r-squared as a reliable result? I mean, is there actually a 90-somthing % fit of the variables? Surely, it is hard to answer without any further kowledge about my data set, etc. but I just found it a bit strange that the r-squared increased that much.
Thank you for any help!
Best regards
Student...
I am currently performing a time-series regression analysis, and need to adjust for serial correlation. I have added "ar(1)" to the end of my equation, and obtained a DW(durbin watson) value that is acceptable. However my R-squared jumped from 0.65 to 0.9-something. Can I use this r-squared as a reliable result? I mean, is there actually a 90-somthing % fit of the variables? Surely, it is hard to answer without any further kowledge about my data set, etc. but I just found it a bit strange that the r-squared increased that much.
Thank you for any help!
Best regards
Student...