Predict
overview      how to     example     data     see also 

Stat > DOE > Factorial > Predict

Use a factorial design model to:

·    predict the value of the response variable (fitted values) for the combinations of variable settings that you request

·    create a confidence interval for the mean response at these combinations

·    create a prediction interval for a single observation at these combinations

·    store the prediction information in the worksheet

See Stored Model Overview for a discussion about how to use stored models.

Dialog box items

Response: Choose a response variable from the drop-down list.

Include covariates in prediction: Check to include covariate values in the predictions.

If your model covariates but you do not include them in the prediction, Minitab averages the prediction over the covariates.

In the second drop-down list, choose whether you will enter variable settings or enter worksheet columns that contain the variable settings.

Enter individual values: Choose to enter variable settings.

Enter columns of values: Choose to enter columns from the worksheet.

Variables table

This table contains a column for each variable in the model and for weights if weights are in the model. Depending on your choice in the second drop-down list, you enter either values or columns.

·    Values - Enter at least one value for each variable. The table columns must be the same length. You must enter numeric or date/time values for continuous variables and weights. For categorical variables, choose a setting from the drop-down.

·    Columns - Enter one column for each variable and for weights. Each column must contain the values for one variable. All columns must be the same length. For a continuous variable, the column must contain numeric or date/time data. For a categorical variable, the column must contain values that match the values you used to fit the model.

<Options>

<Results>

<Storage>

<View Model>