Overview for Predict

Use Predict when you have a model and want to predict the value of the response for the combinations of variable settings that you specify.

For example, real estate appraisers want to predict the sales price of an apartment that has specific characteristics. The appraisers previously fit a model to describe how the sales price of urban apartments relates to several predictor variables including the square footage, the number of available units, the age of the building, and the distance from the city center.

This analysis uses a model that you fit and that Minitab stores. For more information, go to Stored model overview.

Where to find this analysis

If you use the Predictive Analytics Module to create a Linear Regression Model or a Binary Logistic Regression model, then select the analysis from the top of the results. If you create the model from the Stat menu, then use the version of this analysis that corresponds to the type of model you fit.

Type of model	Version of Predict
Regression	Stat > Regression > Regression > Predict
Stability study	Stat > Regression > Stability Study > Predict
Binary logistic regression	Stat > Regression > Binary Logistic Regression > Predict
Poisson regression	Stat > Regression > Poisson Regression > Predict
General linear model	Stat > ANOVA > General Linear Model > Predict
Mixed effects model	Stat > ANOVA > Mixed Effects Model > Fit Mixed Effects Model
Screening design	Stat > DOE > Screening > Predict
Factorial design	Stat > DOE > Factorial > Predict
Response surface design	Stat > DOE > Response Surface > Predict

When to use an alternate analysis

If you have a stored model and want to plot the main effects and interaction effects with fitted means, use Factorial Plots.
If you have a stored model and want to plot the relationship between a fitted response and two continuous variables with contour lines in a two-dimensional view, use Contour Plot.
If you have a stored model and want to plot the relationship between a fitted response and two continuous variables with a three-dimensional response surface, use Surface Plot.
If you have at least one stored model and want to identify an area where the predicted means of one or more response variables are in an acceptable range, use Overlaid Contour Plot.
If you have at least one stored model and want to find values that optimize one or more responses, use Response Optimizer.