Analyze Mixture Design

Mixture Regression
Regression Table - R-Squared

  

Rimage\squared.gif represents the proportion of variation in the response that is explained by the model. Minitab displays three types of Rimage\squared.gif:

·    Rimage\squared.gif (R-Sq) describes the amount of variation in the observed responses that is explained by the model.

·    Adjusted Rimage\squared.gif is a modified Rimage\squared.gif that has been adjusted for the number of terms in the model. If you include unnecessary terms, Rimage\squared.gif can be artificially high. Unlike Rimage\squared.gif, adjusted Rimage\squared.gif may get smaller when you add terms to the model.

·    The predicted Rimage\squared.gif (R-Sq(pred)) reflects how well the model will predict future data.

Note

You can make Rimage\squared.gif artificially high by including too many terms in the regression model. If you add unnecessary predictors to the model, the Rimage\squared.gif usually increases even if you gain no additional information about the response. Because the adjusted Rimage\squared.gif takes into consideration the number of predictors in the model, it is more appropriate than Rimage\squared.gif for comparing models with different numbers of predictors.

Example Output

S = 1.48597     PRESS = 484.810

R-Sq = 99.64%  R-Sq(pred) = 86.92%  R-Sq(adj) = 98.99%

Interpretation

For the fondue data, Rimage\squared.gif is 99.64%, adjusted Rimage\squared.gif is 98.99%, and the predicted Rimage\squared.gif is 86.92%.