Stored model overview

In This Topic

Project models and worksheet models
Analyses that use stored models

Minitab stores models as project models or worksheet models. Both types of models let you quickly and easily create graphs, make predictions, and optimize responses.

Project models and worksheet models

The model types differ in where you access them and how Minitab saves them.

Project models: You access a project model from the toolbar at the top of the results for the model. A project model is active anytime the output for the model is active. Minitab stores project models in the project file (*.MPX). All models from the Predictive Analytics Module are project models.
Worksheet models: You access a worksheet model from the menu of the original analysis. A worksheet model is active when the worksheet that contains the model is active. Minitab stores worksheet models in the worksheet file (*.MWX). All models from the Stat menu are worksheet models.

Where to find details for worksheet models

For project models, the details about a model are in the results that are visible when you work with the model. To find details for worksheet models when the results are not visible, use the following methods.

Model details: After you fit a model for a response variable, an indicator appears in the bottom right corner of the column ID cell on the worksheet. A green indicator tells you that the model is up-to-date while a red indicator signifies an out-of-date model. A model is out-of-date when the data have changed since you fit the model. If you hover the pointer over the indicator, a tool tip appears that provides information about the model.
View Model subdialog box: Each model-based analysis has a View Model button in the main dialog box that you can click before you use the analysis. You can examine all of the models of the same type that are stored in the worksheet. You can only use an analysis if the model is both up-to-date and satisfies the minimum requirements.
View Model Used for Graph: Double-click the graph. Then, right-click the graph and choose View Model Used for Graph from the context menu to see which model created the graph. If a graph is out-of-date, it could be due to changes in the model, data, or worksheet since you created the graph.

Analyses that use stored models

After Minitab stores a model, you can use the following analyses. These analyses are active when the stored model satisfies the minimum requirements.

Analysis	Purpose	Data	Model types
Predict	To predict the value of the response variable (fitted values) for the combinations of variable settings that you request	Any number and type of variables	All worksheet models and all project models
Factorial Plots	To visualize the main effects and interaction effects	Any number and type of variables	All worksheet models Project models for Linear Regression or Binary Logistic Regression
Contour Plot	To show how the fitted response relates to two continuous variables with a two-dimensional view	Two or more continuous variables	Worksheet models, except for Fit Mixed Effects Model Project models for Linear Regression or Binary Logistic Regression
Surface Plot	To show how the fitted response relates to two continuous variables with a three-dimensional surface	Two or more continuous variables	Worksheet models, except for Fit Mixed Effects Model Project models for Linear Regression or Binary Logistic Regression
Cube Plot	To see the factors and combination of settings used in your design	From 2 to 8 factors	Worksheet models for 2-level factorial designs from Analyze Factorial Design
Overlaid Contour Plot	To display contour plots for multiple responses in a single graph to visually identify an area where all of the responses are in a desirable range	Two or more continuous variables	Worksheet models, except for Fit Mixed Effects Model
Response Optimizer	To identify the combination of input variable settings that jointly optimize a single response or a set of responses	Any number and type of variables	Worksheet models, except for Fit Mixed Effects Model Project models for Linear Regression or Binary Logistic Regression
Comparisons	To examine which factor level means are significantly different and to estimate by how much they are different.	At least one factor	Worksheet models for Fit General Linear Model and Fit Mixed Effects Model

Analyses that differ by model type

Prediction in different worksheets: A project model is available no matter what worksheet is active, so you can predict for columns of data that are in a different worksheet from the response variable. A worksheet model is available only when the worksheet for the model is active. To predict from columns of data for a worksheet model, add columns to the same worksheet as the response.
Multiple response optimization: A project model has access to only the response variable in the results for the model. Project models let you optimize only the response in the results. For multiple response optimization, add all the response variables to the same worksheet and fit a worksheet model for each response.