In: Statistics and Probability
Which is more appropriate in evaluating a model?
R-squared = 51.17 %
R-squared (adjusted for d.f.) = 50.09 %
The given information is:
R-squared = 51.17 %
R-squared (adjusted for d.f.) = 50.09 %
R-squared also referred as coefficient of determination.
R-squared explains the proportion or the total amount of deviation in the response variable that can be explained by the every explanatory variables involved in the statistical model whereas adjusted R-squared explains the proportion or the total amount of variation in the response variable that can be explained by those explanatory variables that really affects the statistical model.
Both the measures explains the accuracy of the statistical model. But adjusted R-squared tells about the explanatory variables that are useful or not useful to increase the accuracy. Adjusted R-squared explains what will be the effect on model if we add or remove the variables from the model that are not useful.
Therefore, it can be said that Adjusted R-squared is more more appropriate in evaluating a model.