Hi Andy, It is because it loses the useful interpretation it has for linear regression. That is, it is no longer the variance explained by the model. It is rather a measure that includes the MSE scaled by the variance.
There might be some contexts where this might be useful. However, on one hand we lose the a sense of the unit of measurement using this metric (which not the case for example for metrics such as MAE). On the other hand, the correlation between true and predicted values can be a quite simpler and clearer measure to use (even though this can, of course, be a matter of preference). The concordance correlation coefficient builds on top of that. 2015-09-07 20:38 GMT+01:00 Andy <t3k...@gmail.com>: > On 09/07/2015 06:03 AM, Stylianos Kampakis wrote: > > > > The interpretation of R^2 is less useful for machine learning models. > > For example, Weka omits it all together for regression models. A > > useful alternative is to simply use the correlation between the true > > and the predicted values. > Can you explain this? > Why do you think it is not useful? > > > ------------------------------------------------------------------------------ > _______________________________________________ > Scikit-learn-general mailing list > Scikit-learn-general@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general >
------------------------------------------------------------------------------
_______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
