Run
Select an Alternative Model
button after the Model Selection table. . Click the

This command is available with the Predictive Analytics Module. Click here for more information about how to activate the module.

When you use Discover Best Model (Binary Response) to identify the best model type, Minitab Statistical Software produces results for the model with the best value of the accuracy criterion for the analysis, such as the minimum average –loglikelihood. Minitab lets you explore results for other models and other types of models. For example, if another type of model produces similar prediction accuracy, you can determine whether the same predictors are important in each type of model.

The available options depend on the type of model. For CART^{®} and
binary logistic regression models, you can examine the results for the best
model from the search. For Random Forests^{®} and TreeNet^{®}
models, you can examine results from any of the models in the search. For
Random Forests^{®} and TreeNet^{®} models, you can also tune
the hyperparameters to look for combinations that produce even better values
than the hyperparameters in the search.

Select an existing model to produce results for one of the models from the search. Specify hyperparameters to fit new models to look for combinations of hyperparameters that improve the performance of the model.

In the search for the best type of model, the analysis produces up to 3
Random Forests^{®} models with different minimum sizes for internal
nodes. Select a model from the list and click
Display
Results
to produce results for that model.

The analysis requires that you specify the first 3 hyperparameters. The
inclusion of a bootstrap sample size less than the training data size is
optional. Click
Display
Results
to evaluate the hyperparameters for the new models. The results include a table
that compares the optimality criteria for the different combinations of
hyperparameters and the model results for the model with the best value of the
optimality criterion, such as the minimum average –loglikelihood.

- Number of predictors for node splitting
- Specify 1 to 3 numbers of predictors to consider for each node split. Typically, the analysis works well when you consider the square root of the total number of predictors. However, some data sets have associations among the predictors that lead to improved model performance when the analysis considers a larger or smaller number of predictors for each node.
- Minimum number of cases to split an internal node
- Enter 1 to 3 minimum numbers of cases a node can have and still split into more nodes. By default, the original search includes the numbers 2, 5, and 8.
- Number of bootstrap samples to grow trees
- Enter a value to determine the number of bootstrap samples and the number of trees produced by the analysis. Enter a value between 3 and 3000.
- Specify a bootstrap sample size less than the training data size
- Select to enter a value that sets the bootstrap sample size. You must enter a value greater than or equal to 5. If you enter a size that is greater than the training data size, Minitab uses a sample size equal to the training data size.

Select an existing model to produce results for one of the models from the search. Specify hyperparameters to fit new models to look for combinations of hyperparameters that improve the performance of the model.

In the search for the best type of model, the analysis produces a
TreeNet^{®} model for each combination of hyperparameters. Select a
model from the list and click
Display
Results
to produce results for that model.

The analysis requires that you specify all of the hyperparameters. Click Display Results to evaluate the hyperparameters for the new models. The results include a table that compares the optimality criteria for the different combinations of hyperparameters and the results for the model with the best value of the accuracy criterion for the analysis, such as the minimum average –loglikelihood.

- Learning rate
- Enter up to 10 values. Eligible values are from 0.0001 to 1.
- Subsample fraction
- Enter up to 10 values. Eligible values are greater than 0 and less than or equal to 1.
- Maximum terminal nodes per tree and Maximum tree depth
- Choose whether to evaluate the
Maximum
terminal nodes per tree
or the
Maximum tree
depth.
Usually, either choice is a reasonable way to identify a useful model, so that
the selection depends only on individual preference.
- Maximum terminal nodes
- Enter up to 3 values. Eligible values are between 2 and 2000. A value of 2 eliminates the investigation of interactions.
- Maximum tree depth
- Enter up to 3 values. Eligible values are between 2 and 1000 to represent the maximum depth of a tree. The root node corresponds to a depth of 1. In many applications, depths from 4 to 6 give reasonably good models

- Number of predictors for node splitting
- Enter up to 3 values. Eligible values are between 1 and the total number of predictors. Usually, the analysis works well when you consider the total number of predictors. However, some data sets have associations among the predictors that lead to improved model performance when the analysis considers a smaller number of predictors for each node.
- Number of trees
- Enter a value between 1 and 5000 to specify the maximum number of trees to build. The default value of 300 usually provides useful results for the evaluation of the hyperparameter values.

Select Results for logistic regression model and click Display Results to produce the results for the best binary logistic regression model from the search for the best type of model.

Select
Results for CART® model
and click
Display
Results
to produce the results for the best CART^{®} model from the search for
the best type of model.