Model Profiler
Following are the ways to access the Model Profiler:
In the Distributions & Simulation Startup Panel, select Profiler and click the OK button to display the Model Profiler.
In the GLM Results dialog box, on the Report tab, click the Model profiler button to display the Model Profiler.
In the GRM Results dialog box, on the Report tab, click the Model profiler button to display the Model Profiler.
In the DOE Analysis Results dialog box, on the Model tab, click the Model profiler button to display the Model Profiler.
The Model Profiler contains two tabs: Profiler and Output. The following options are available regardless of which tab is selected.
Simulate. Click this button to simulate the responses and create the predicted profiles graph and the simulated spreadsheet. Additionally, the defect rate spreadsheet is created if a valid specification limit is set for any of the dependent variables.
Auto update spreadsheet. When this check box is selected, the spreadsheet produced when the Simulate button is clicked will be auto-updated.
Defect Profiler. This button becomes available when a valid specification limit is set. Click this button to produce the defect profile. The defect profile shows the probability of an out-of-spec defect as a function of each factor, while the other factors vary randomly. This is used to help visualize which factor’s distributional changes the process is most sensitive to, in the quest to improve quality and decrease cost.
Specification limits define what a defect is, and random factors provide the variation to produce defects in the simulation. Both need to be present for a defect profile to be meaningful.
At least one of the factors must be declared random for a defect simulation to be meaningful, otherwise the simulation outputs would be constant. These are specified through the simulator Factor specifications.
Load models. Click this button to display the Open PMML files dialog box, where you can browse to and open the PMML input model files. More than one model file can be loaded, and in this case, all models will be applied to the data specified as the current active data set. This button will generate random values for the predictor variables based on default distributions (normal for continuous and discrete for categorical) and calculate the values for dependent variables. A spreadsheet containing these values and the predicted profiles graphs will be output.