|
|
|
| R2011b Documentation → Model-Based Calibration Toolbox | |
Learn more about Model-Based Calibration Toolbox |
|
|
|
|
| R2011b Documentation → Model-Based Calibration Toolbox | |
Learn more about Model-Based Calibration Toolbox |
|
| Contents | Index |
| On this page… |
|---|
When you select a response feature node or one-stage model node
in the model tree, this view appears. Both kinds
of models have a global icon (
), to reflect the fact that a global model is
fitted, so this is referred to as global level.
Plots shown here are referred to as global plots. Child nodes of these
models also have global icons. Any node with a global icon produces
this view.
For one-stage models, this view shows the functionality available at all model nodes. For two-stage models there are other levels with different functionality for local level and response level models.
This view is similar in format to the local level view, which also contains scatter plots above special plots. The statistical information panes on the right side are different, there is a Removed Data pane, and there are no test number controls. There is an edit box for model comments. See also:
The lower plots in the global level view are referred to as special plots, as they can be different for different models (for example, none at all for neural net models).
The special plot at the global level shows a Predicted/Observed plot. Where there is only one input factor, the plot shows the model fit and the data against the input factor (as in most local model special plots, which often have only one input factor).
For response feature models, each data point is the value taken by this response feature for some local model fit (of this two-stage model). Note that response features are not necessarily coefficients of the local curves, but are always derived from them in some way.
When there is more than one input factor it becomes impossible to display the fit in the same way, so the data for the response feature is plotted against the values predicted by the global model. The line of predicted=observed is shown. With a perfect fit, each point would be exactly on this line. The distances of the points from the line (the residuals) show how well the model fits the data.
To examine the fit in more detail, double-click the arrows (indicated in the figure in Global Models) to hide the scatter plot and expand the lower plot. You can also zoom in on parts of the plot by Shift-click-dragging or middle-click-dragging on the place of interest on the plot. Return to full size by double-clicking.
Note Right-click a point in either the special or scatter plot to open a figure plot of that particular test (for example, torque against spark). |
You can select the plot type from the drop-down menu at the top of the plot. Choices here can be:
Predicted/Observed
Normal Plot Normal plots are a useful graph for assessing whether data comes from a normal distribution. For more information, see Normal Probability Plots in the Statistics Toolbox documentation.
Validation Residuals — This option is only available if you are viewing a one-stage model. If you are using validation data, the plot shows the one-stage model validation residuals. Validation data must be attached at the Test Plan Level. See Using Validation Data.
Surface and Contour are only available for models with two inputs. You can view a surface or contour plot with data points.
The upper plots in the global level view are referred to as scatter plots. They can show various scatter plots of statistics for assessing goodness-of-fit for the current model shown. The upper scatter plots are replaced by an icon if you resize the Browser too small.
The statistics available for plotting are model dependent.
The preceding are the context menus for both plots. On both plots you can manipulate outliers with all the same commands available in the Outliers menu except Outlier Selection Criteria and Copy Outliers. See Outliers Menu (Local Level) for details. For one-stage models you can select Plot Variables on the scatter plot.
You can choose the x- and y-axis factors using the drop-down menus. The available statistics and factors are model dependent. Following is an example.
Shown is an example drop-down menu on the scatter plot for changing x and y factors. In this example knot is the response feature node selected. Therefore the model output is knot, so knot and Predicted knot are available in the menu. (For child nodes of knot, the model output is still knot.) The global inputs, the model output, and the predicted model output are always available in these menus. The observation number is also always available.
The other options are statistics that are model dependent, and can include: Residuals, Weighted Residuals, Studentized Residuals, Leverage, and Cook's Distance. These statistics and the other factors are also used as the available criteria for selection of outliers, so the options in the Outlier Selection Criteria dialog box are similarly model dependent. At global (or one-stage) level these are externally studentized residuals.
Summary Table
Observations | Number of observations used to estimate model |
Parameters | Number of parameters in model |
Box-Cox | Power transform used for box-cox transformation. A value of zero means a log transform is used. A value of 1 means there is no transformation. |
PRESS RMSE | Root mean squared error of predicted errors. The divisor used for PRESS RMSE is the number of observations. The residuals are in untransformed values to enable comparison between alternative models with different Box-Cox transformations. |
RMSE | Root mean squared error. The divisor used for RMSE is the number of observations minus the number of parameters. The residuals are in untransformed values, to enable comparison between alternative models with different Box-Cox transformations. |
Validation RMSE (one-stage models only) | Root mean squared error between the one-stage model and the validation data. See Using Validation Data. |
Select Model > Summary Statistics to change and add to the Summary Statistics displayed in the Summary Table.
ANOVA Table
| SS | df | MS | |
|---|---|---|---|
Regression | SSR | p-1 | SSR/(p-1) |
Error | SSE | n-p | SSE/(N-p) |
Total | SST | n-1 |
This table is not present after you have calculated MLE. See also Toolbox Terms and Statistics Definitions.
Tests you have removed (using the Remove Outliers menu item) are shown in a list. Select Outliers > Restore Removed Data to select some or all of these to restore. Tests marked with an asterisk (*) are not restorable here (at global level). Such tests cannot be restored because the entire test has been removed at the local level (using the Remove All Data menu item), or the local model could not be fitted. Removed tests can only be restored at the local level.
Double-click on any removed test number to display a plot of the test in a figure window.
You can enter comments in the edit box.
Here is a list of any child models of the currently selected global model. This list is empty if there are no child nodes to compare. If there are child nodes you can click the Select button here to enter the Model Selection window to compare the child models and choose the best.
Click New to add a new child global model (or Delete to remove one). For more information see the Test Plans List Pane. The contents of this pane change in different views; it always contains the child nodes of the node selected in the model tree (and the New, Delete, and Select buttons). At the global level it contains a list of any child models.
The list view displays the number of parameters and observations, the value of any Box-Cox transformation (1 indicates no transform), and the values of PRESS RMSE (for linear models only) and RMSE for each child model.
If you are viewing a one-stage model and are using validation data, Validation RMSE is displayed. See Using Validation Data.
Note Note that you can specify additional statistics to display here and in the Summary table by selecting Model > Summary Statistics. See Summary Statistics. |
Use these statistics to compare the fit of different child models and help you choose the best. For definitions of RMSE and PRESS RMSE, see Summary Table and Model Selection Guide. For other statistics see Model Menu (Global Level). For information on Box-Cox transforms, see Box-Cox Transformation.
Note that the Summary Statistics settings are inherited from parent global models and from the test plan level only if set before the current node was created, when they also appear in the bottom list pane if there are child nodes to compare.
This toolbar appears when a response feature node or one-stage model node (both have a global icon) is selected in the model tree. Note that for one-stage models all model child nodes of the one-stage test plan are of this type.
Further buttons appear on the right depending on the type of model at the node selected. See Global Level: Model-Specific Tools.
The eight left icons remain constant throughout the levels. They are for project and node management, Help, and here the print icon is enabled, as there are plots in this view. See Project Level: Toolbar for details on these buttons.
View Model — Opens a dialog box displaying the terms in the current model. See Model Definition Dialog Box.
Box-Cox Transform — Opens the Box-Cox Transformation plots, where you can minimize SSE to try to improve the fit. See Box-Cox Transformation for statistical details.
Build Models — Opens the Build Models Dialog Box, where you can choose a template for the type of models you want to build. There are predefined templates for polynomials, RBFs and hybrid RBFs, and free knot splines, you can create new templates, and you can use any suitable parent node in the current project as a template. You can also save templates of whatever models you choose using the Model > Make Template menu item. User-defined templates can then be found via the Build Models dialog box. You can use the Browse button to find stored templates that are not in the default directory.
View Modeling Data — Opens the Data Editor, where you can view a read-only version of the inputs, predicted and actual responses. This allows you to examine and export your modeling data, using all the powerful display features of the Data Editor. Also in the View menu.
These four toolbar icons appear for every global model node (although Box-Cox is not enabled for neural net models). The icons that appear to the right are model specific.
All twelve left buttons (up to View Modeling Data) appear for all response feature models and one-stage models. See Global Level: Toolbar for details on these buttons. The right buttons change according to model type.
Stepwise — This opens the Stepwise Regression window, where you can view the effects of removing and restoring model terms on the PRESS statistic (Predicted Error Sum of Squares), which is a measure of the predictive quality of a model. You can also use Min PRESS to remove all at once model terms that do not improve the predictive qualities of the model. See Stepwise for further discussion of the statistical effects of the Stepwise feature.
Design Evaluation — Opens the Design Evaluation tool, where you can view properties of the design. See Design Evaluation Tool.
Prediction Error Variance Viewer - Opens the Prediction Error Variance Viewer. See Prediction Error Variance Viewer.
Free knot spline models do not have any model-specific tools, just the standard View Model, Box-Cox Transform, Build Models and View Modeling Data.
Update Fit refits the RBF widths and centers. See Tips for Modeling with Radial Basis Functions and Fitting Routines in the Radial Basis Functions chapter.
View Centers opens a dialog box where you can view the position of the radial basis function's centers graphically and in table form.
Prune opens the Number of Centers Selector where you can minimize various error statistics by decreasing the number of centers. See Prune Functionality.
Stepwise opens the Stepwise Regression window.
Design Evaluation - Opens the Design Evaluation tool, where you can view properties of the design. See Design Evaluation Tool.
Prediction Error Variance Viewer - Opens the Prediction Error Variance Viewer.
Hybrid RBFs have the same toolbar buttons as linear models.
This toolbar appears when you select any response feature that is an MLE model (purple icon). See Global Models for other functionality in this view.
At this point the New child node, Box-Cox and Build Models buttons are disabled.
Recalculate MLE returns to the MLE dialog box, where you can perform more iterations to try to refine the MLE model fit. See MLE for more details.
Neural net models have the View Model, Build Models, and Update Fit tools.
Only the New (child node) and Delete (current node) functions change according to the node level currently selected. Otherwise the File menu remains constant.
See File Menu.
The Window and Help menus have the same form throughout the Model Browser.
See Window Menu and Help Menu.
Set Up opens the Global Model Setup dialog box, where you can change the model type. See Global Model Setup.
Reset opens a confirmation dialog box so you cannot unintentionally reset your model. When you confirm you want to continue, the model is reset to the global model default, that is, the global model specified at the test plan stage, restoring any removed outliers and removing any transforms.
Summary Statistics opens the Summary Statistics dialog box where you can select which statistics to display to help you evaluate models. In the global level view these appear in the Summary Table to the right of the plots, and in the Models list pane if there are child node models to compare. They also appear in the Model Selection window, and can be used to automatically select the best child node when using the Build Models dialog box and Local Multiple Models. The standard summary statistics are PRESS RMSE (for linear models only) and RMSE, and these are always displayed. You can choose additional statistics for display in the Summary Statistics dialog box by selecting the check boxes. You can also reach the dialog box from the test plan. When you create child nodes the selected summary statistics are inherited. See Summary Statistics for more information.
Evaluate — Select from the submenu Fit Data, Validation Data, No Data, or Other Data. Opens the Model Evaluation window displaying the fit relative to the selected data (or no data). You must attach validation data to the test plan before you can use it. If you select Other Data a wizard appears to select the data.
Box-Cox Transform opens the Box-Cox Transformation plots, where you can minimize SSE to try to improve the fit. See Box-Cox Transformation for statistical details.
Utilities opens a submenu showing the model-specific options available, duplicating the model-specific toolbar buttons (for example, Stepwise, Design Evaluation, View Centers, Prediction Error Variance Viewer, and so on).
Make Template is available when child nodes exist. This opens a file browser where you can choose to save all the current child node models as a template, which you can then access using the Build Models menu item or toolbar button.
Build Models opens the Build Models dialog box. Here you can create a template, use a predefined template, or use current models as a template to build a selection of models as child nodes of the current node (or for local models, as child nodes of each response feature). The best model of each selection of child nodes will be automatically selected (it will have a blue icon), based on the selection criteria you choose in the following dialog box (such as PRESS RMSE, RMSE, Box-Cox, Observations or Parameters).
See Build Models Dialog Box for details.
Select is available whenever the Select button is also enabled in the lower right pane (when it is titled Local Models, Response Features, or Models). This item opens the Model Selection window to allow you to choose the best model. See Select Button.
Assign Best selects the current model as best. If it is one of several child node models of a global model, selecting it as best means that it is duplicated at the parent global model. See Model Tree.
Model Definition opens the Model Definition dialog box displaying the model terms. See Model Definition Dialog Box.
Modeling Data opens the Data Editor to display a read only version of the inputs and predicted and actual responses. Here you can view and export your modeling data. Also in the toolbar.
Test Numbers turns test numbers on and off for both the special and scatter plots. Also available in the right-click plot menus.
Model Definition Dialog Box. You can open this dialog box for any model to view the parameters and coefficients of the model formula and the coding details. You can view the Model Definition dialog box at the local, global, and response levels. For any radial basis function model you can see the kernel type, number of centers, width, and regularization parameter, as shown in the following example.
For radial basis function models, see also How to Find RBF Model Formula.
This is the same as the local level Outliers menu, except that there is no Remove All Data command. All items are duplicated in the right-click context menu on the plots, except Selection Criteria and Copy Outliers. See Outliers Menu (Local Level).
At global level, as at local level, the Restore Removed Data item opens the Restore Removed Data dialog box, where you can choose the points to restore from the list, or restore all available points. Select points in the left list and use the buttons to move points between the lists. Note that entire tests removed at the local level (using the Remove All Data item) cannot be restored at global level.
 | Local Models | Selecting Models |  |
Includes the most popular MATLAB recorded presentations with Q&A sessions led by MATLAB experts.
| © 1984-2012- The MathWorks, Inc. - Site Help - Patents - Trademarks - Privacy Policy - Preventing Piracy - RSS |
