Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.


Chun
Posts:
2
From:
USA
Registered:
11/23/09


Re: Negative R^2 value
Posted:
Nov 23, 2009 10:48 PM


The coefficient of determination is always in [0,1] for linear regression. For statistical procedures other than linear regression, ad hoc "Rsquared type" measures exist that try to mimic the behavior of Rsquared, but may end up being negative. The "adjusted" Rsquared can also be negative.



