Partial Least Squares -
Optionsmain topic how to see also
Partial Least Squares -
OptionsStat > Regression > Partial Least Squares > Options
You can use cross-validation to select the components that maximize your model's predictive ability, to specify the coding scheme used for categorical predictors, and to specify the reference level.
Cross-Validation
None: Choose to suppress cross-validation.
Leave-one-out: Choose to perform cross-validation leaving out one observation each time the model is recalculated.
Leave-group-out of size: Choose to perform cross-validation leaving out multiple observations (groups) each time the model is recalculated. Then, enter group size, which is 2 by default.
Leave out as specified in column: Choose to perform cross-validation using group identifiers (positive integers) to specify which observations are left out together each time the model is recalculated. Then, enter group identifier column, which must be equal in length to predictor and response columns and contain no missing values.
Type of coding for categorical predictors: Use when you have categorical predictors. To perform the analysis, Minitab needs to recode the categorical data. See Coding categorical predictors in Partial Least Squares for more information.
(1, 0): Choose if you want categorical predictors coded as 1, 0.
(- 1, 0, +1): Choose if you want categorical predictors coded as - 1, 0, 1.
Reference level (enter categorical predictor followed by level): Enter the reference level by typing the categorical predictor column followed by the reference level. (Text and date/time levels must be enclosed in quotes.) You can assign a reference level only when you use 1, 0 coding.