Model Fit Report

In the Partial Least Squares platform, the Model Fit Report shows detailed results for each fitted model. The fit uses either the optimum number of factors based on cross validation, or the specified number of factors if no cross validation methods are specified. The report title indicates whether NIPALS or SIMPLS was used, whether Fast SVD or Classical SVD was used, and gives the number of extracted factors.

Figure 6.12 Model Fit Report 

The Model Fit report includes the following summary information:

X-Y Scores Plots

Scatterplots of the X and Y scores for each extracted factor.

Percent Variation Explained

Shows the percent variation and cumulative percent variation explained for both X and Y. Results are given for each extracted factor.

Crossvalidation

(Available only when a validation column or random holdback method is used.) Shows a table of model performance statistics for the training and validation sets. For each set, the table includes the coefficient of determination (Rsquare), the root average squared error (RASE), and the number of observations used (Freq). These statistics enable you to compare model fit and predictive accuracy across data partitions and assess how well the model generalizes to new data.

Model Coefficients for Centered and Scaled Data

For each Y, shows the coefficients of the Xs for the model based on the centered and scaled data.