Deployment Model - Results Tab

Select the Results tab of the Deployment Model dialog box to access options to execute the active model. Active models can be selected from the Model drop-down list as described below.

NOTE: For specific details on Hotelling T2, SPE (Q), and other technical functions mentioned in the option descriptions below, see PCA and PLS Technical Notes.

Element Name	Description
Model	Use this drop-down list to select an active model for generating spreadsheets and graphs. Although you can have multiple model files loaded, you can only execute one model at a time against the currently active data set.
Model summary	Click this button to display a spreadsheet listing the names of the loaded PMML files and their status (whether they have been successfully loaded), as well as the number of their variables and components.
Results	The following eight options are displayed in the Results group box:
X	Select this option button to generate spreadsheets of predictions, residuals, scores, and loadings for the X (dependent) variables. PCA and PLS (and their batch versions such as BMPLS, TMPCA, and TMPLS) transform the representation of the predictor variables into a new set of coordinates that captures, at least in principle, the maximum of the variability in the original data. Therefore, these models are capable of predicting the independent variables, just as a regression model is capable of predicting the dependent variables. In addition, PLS can model a set of dependent variables Y based on the predictors X. In other words, PLS can perform both regression and classification tasks.
Y	Available only for PLS and TMPLS models. Select this option button to generate spreadsheets of predictions, residuals, loadings, etc, for the dependent Y variables.
Predictions	Click this button to generate predictions of the active model for the X or Y variables in spreadsheet format.
Residuals	Click this button to generate a spreadsheet of residuals for the X or Y variables. Residuals are differences between the original variables and their predicted values made by the active model. In other words, it is that part of variability in the data that could not be explained by the model.
Loadings	Click this button to generate a spreadsheet of loading factors for the X or Y variables. The loading factors are coefficients that completely determine the orientation of the principal components with respect to the original coordinate system. In other words, a loading factor for a variable in the direction of a principal component indicates the importance of that variable in determining that component.
Scores	Click this button to generate a spreadsheet of t-scores (for X) or u-scores (for Y) (see PCA and PLS Technical Notes) along each principal component. Scores are the representation of the original variables in the new coordinate system defined by the principal components. Thus corresponding to each data case, say, x = (x1, x2, …), is a set of score values, t = (t1, t2, …), which uniquely maps a point in the variable space onto a point in the principal space.
Save scores	Click this button to display a standard variable selection dialog, which is used to select variable(s) to be displayed together with the scores (for X or Y). After you select the variable(s), a spreadsheet containing the specified variable(s) will be displayed in an individual window, regardless of the settings on the Options dialog box - Output Manager tab or the Analysis/Graph Output Manager dialog box. You can, however, add the spreadsheet to a workbook or report using the or buttons, respectively. Note: To save the spreadsheet,  select the spreadsheet, then select Save or Save As from the File menu. Saving is useful if you want to use the residual values for further study with other Statistica analyses.
Scores	Use the options in this group box to generate results in the form of t-scores (X), u-scores (Y), and t control charts. Further options for generating variable contributions and brushing applications are also provided.
Component	On this drop-down menu, select a principal component for which you want to generate contributions and scores.
Score (t)	Click this button to generate a t-scores (X) lineplot along the direction of the specified principal component (see the Component option description above). The t-scores are the representation of the X variables in the new coordinate system, defined by the principal components.
(Score [t] Brushing)	Click this button to generate a t-scores lineplot (see the previous description) and a Brushing command dialog box.
(Score [t] Spreadsheet)	Click this button to generate a t-scores spreadsheet. See the Score (t) option description above for more details.
Score (u)	Click this button to generate a u-scores (Y) lineplot along the direction of the specified principal component (see the Component option description above). The u-scores are the representation of the Y variables in the new coordinate system defined by the principal components. Available only for PLS and TMPLS models that have an explicit time dependency.
(Score[u] Brushing)	Click this button to generate a u-scores lineplot (see the previous description) and a Brushing command dialog box.
(Score [u] Spreadsheet)	Click this button to generate a u-scores spreadsheet. See the Score (u) option description above for more details. They are available only for models such as PLS and TMPLS with explicit time dependency.
Control (t)	Click this button to generate a control lineplot for the t-scores, along the direction of the specified principal component (see the Component option description above). The t-scores are the representation of the X variables in the new coordinate system, defined by the principal components. This option is available only for BMPLS models.
(Control [t] Brushing}	Click this button to generate a t-scores control chart (see the previous description) and a Brushing command dialog box.
(Control [t] Spreadsheet)	Click this button to generate a control spreadsheet for the t-scores along the direction of the specified principal component (see the Component option description above). The t-scores are the representation of the X variables in the new coordinate system defined by the principal components. This option is available only for BMPLS models.
Contribution	Use the options in this group box to generate contribution outputs in the form of spreadsheets and graphs.
Contribution	Click this button to generate a histogram of variable contributions to the t-scores (X) for a specified observation along the direction of a specified principal component (see the Component option description above) Variables with axes aligned with the direction of a principal component make stronger contributions to the scores. For PCA and PLS models (non-batch), an observation in the data set is uniquely determined by its case number (see the Obs. option description below). For batch data sets, an observation is identified by its time slot and batch membership.
(Contribution Spreadsheet)	Click this button to generate a spreadsheet of variable contributions to the t-scores for a specified observation along the direction of a principal component (see the Component option description above).
Obs	(for PCA and PLS models) (Batch for BMPLS, TMPCA, and TMPLS models) This option can be called Obs. or Batch. Use this option to identify an observation in the data set. For PCA and PLS models (i.e., non-batch), an observation in the data set is uniquely determined by its case number. For batch data sets, an observation is identified by its time slot and batch membership. In other words, a batch observation in a batch data set can only be fully identified by its batch membership and the time slot in which it was measured. See the Time option description below for more details.
Time	Using this option, together with Batch (see above), you can uniquely identify an observation in a batch data set. It is available only for BMPLS, TMPCA, and TMPLS models that have an explicit time dependency.
T2 Chart	Click this button to generate line plots of the Hotelling T2 and the upper control limit (99%) for the observations in the active data set. T2 plays an important role in case wise diagnostics of data. These plots are particularly suitable for detecting outliers that are recognized by their relatively large value of T2.
(T2 Spreadsheet)	Click this button to generate a spreadsheet of the Hotelling T2 scores and the upper control limit for the observations in the active data set.
SPE(Q) Chart	Click this button to generate an SPE (Q) chart in histogram format. SPE represents noise (unstructured variation) in the data that cannot be captured by the model, since it does not follow a particular pattern.
(SPE[Q] Spreadsheet)	Click this button to generate the same observations as described above in spreadsheet format.
D-To-Model	Click this button to display a line plot of distance-to-model for the observations in the data set. Distance-to-model plays an important role in process control since it measures the squared perpendicular distance of an observation from the normal plane. Distance-to-model is used as an indication of whether an observation is within the domain of normality.  Hence, it can be used for detecting outliers.
(D-To-Model  Spreadsheet)	Click this button to generate the same observations as described above in spreadsheet format.
Variable	Use the options in this group box to generate variable trajectories.
Trajectory	Click this button to generate a graph for the trajectories of the variable specified in the variable drop-down-list (described below) for all batches in the data set. Trajectories record the development of measurement variables during the evolution of a batch. A trajectory graph contains as many plots as the number of batches in the data set.
(Trajectory Spreadsheet)	Click this button to generate the same data in spreadsheet format.
Variable drop-down-list	From this list, select the variable for which you want to generate a trajectory plot or spreadsheet.
Limits as regions	Select this checkbox to display the area inside the control and warning limits of the variable trajectory chart as different colors in order to more easily identify the in-control and out-of-control batches.