The first time you click Go in the Model Launch control panel (Partial Least Squares Model Launch Control Panel), the Validation Method panel is removed from the Model Launch window. If you specified a Validation column or if you selected Holdback in the Validation Method panel, all model fits in the report are based on the training data. Otherwise, all model fits are based on the entire data set.
If you selected None as the CV method, two reports appear:
In Model Comparison Summary, models for 7 and then 6 factors have been fit. The report includes the following summary information:
This report appears only when a form of cross validation is selected as a Validation Method in the Model Launch control panel. It shows summary statistics for models fit, using from 0 to the maximum number of extracted factors, as specified in the Model Launch control panel. The report also provides a plot of Root Mean PRESS values. See Root Mean PRESS Plot. An optimum number of factors is identified using the minimum Root Mean PRESS statistic.
When the Standardize X option is selected, cross validation is applied once to the entire data table. It is not reapplied to the individual training sets. However, when any combination of the Centering or Scaling options are selected, this combination of selections is applied to each cross validation training set. Cross validation proceeds by using the training sets, which are individually centered and scaled if these options are selected.
Test statistic for the van der Voet test, which tests whether models with different numbers of extracted factors differ significantly from the optimum model. The null hypothesis for each van der Voet T2 test states that the model based on the corresponding number of factors does not differ from the optimum model. For more details, see van der Voet T2.
Q2
Indicator of the predictive ability of models with the given number of factors or fewer. For a given number of factors, f, Cumulative Q2 is defined as follows:
R2X
Percent of X variation explained by the model with the given number of factors. See Calculation of R2X and R2Y When Validation Is Used.
R2Y
Percent of Y variation explained by the model with the given number of factors. See Calculation of R2X and R2Y When Validation Is Used.
This bar chart shows the number of factors along the horizontal axis and the Root Mean PRESS values on the vertical axis. It is equivalent to the horizontal bar chart that appears to the right of the Root Mean PRESS column in the Cross Validation report. See Cross Validation Report.
For a specified number of factors, a, Root Mean PRESS is calculated as follows:
1.
|
3.
|
4.
|
Root Mean PRESS for a factors is the square root of the average of the PRESS values across all responses.
|
The statistic Q2 is defined as . The PRESS statistic is the predicted error sum of squares across all responses for the model developed based on training data, but evaluated on the validation set. The value of SSY is the sum of squares for Y across all responses based on the observations in the validation set.
The statistic Q2 in the Cross Validation report is computed in the following ways, depending on the selected Validation Method:
Q2 is the average of the values computed for the validation sets based on the models constructed by leaving out one observation at a time.
Q2 is the average of the values computed for the validation sets based on the K models constructed by leaving out each of the K folds.
Q2 is the value of computed for the validation set based on the model constructed using the single set of training data.
The statistics R2X and R2Y in the Cross Validation report are computed in the following ways, depending on the selected Validation Method:
R2X is the average of the Percent Variation Explained for X Effects for the models constructed by leaving out one observation at a time.
R2X is the average of the Percent Variation Explained for X Effects for the K models constructed by leaving out each fold.
R2X is the Percent Variation Explained for X Effects for the model constructed using the training data.