Classification Stump¶

A Classification Decision Stump is a model that consists of a one-level decision tree where the root is connected to terminal nodes (leaves) [Friedman2017]. The library only supports stumps with two leaves. Two methods of split criterion are available: gini and information gain. See Classification Decision Tree for details.

Batch Processing¶

A classification stump follows the general workflow described in Classification Usage Model.

Training¶

For a description of the input and output, refer to Classification Usage Model.

At the training stage, a classification decision stump has the following parameters:

Training Parameters for Classification Stump (Batch Processing)¶
Parameter	Default Value	Description
`algorithmFPType`	`float`	The floating-point type that the algorithm uses for intermediate computations. Can be `float` or `double`.
`method`	`defaultDense`	Performance-oriented computation method, the only method supported by the algorithm.
`splitCriterion`	`decision_tree::classification::gini`	Split criteria for classification stump. Two split criterion are available: `decision_tree::classification::gini` `decision_tree::classification::infoGain` See Classification Decision Tree chapter for details.
`varImportance`	`none`	Note Variable importance computation is not supported for current version of the library.
`nClasses`	\(2\)	The number of classes.

Prediction¶

For a description of the input and output, refer to Classification Usage Model.

At the prediction stage, a classification stump has the following parameters:

Training Parameters for Classification Stump (Batch Processing)¶
Parameter	Default Value	Description
`algorithmFPType`	`float`	The floating-point type that the algorithm uses for intermediate computations. Can be `float` or `double`.
`method`	`defaultDense`	Performance-oriented computation method, the only method supported by the algorithm.
`nClasses`	\(2\)	The number of classes.
`resultsToEvaluate`	`classifier::computeClassLabels`	The form of computed result: `classifier::computeClassLabels` – the result contains the `NumericTable` of size \(n \times 1\) with predicted labels `classifier::computeClassProbabilities` – the result contains the `NumericTable` of size \(n \times \text{nClasses}\) with probabilities to belong to each class

Examples¶

Batch Processing:

oneDAL documentation