Skip to content

ExhaustiveSplitter

Numeric attribute observer for classification tasks that is based on a Binary Search Tree.

This algorithm1 is also referred to as exhaustive attribute observer, since it ends up storing all the observations between split attempts2.

This splitter cannot perform probability density estimations, so it does not work well when coupled with tree leaves using naive bayes models.

Attributes

  • is_numeric

    Determine whether or not the splitter works with numerical features.

  • is_target_class

    Check on which kind of learning task the splitter is designed to work. If True, the splitter works with classification trees, otherwise it is designed for regression trees.

Methods

best_evaluated_split_suggestion

Get the best split suggestion given a criterion and the target's statistics.

Parameters

  • criterion (river.tree._split_criterion.base_split_criterion.SplitCriterion)
  • pre_split_dist (Union[List, Dict])
  • att_idx (Hashable)
  • binary_only (bool)

Returns

SplitSuggestion: Suggestion of the best attribute split.

clone

Return a fresh estimator with the same parameters.

The clone has the same parameters but has not been updated with any data. This works by looking at the parameters from the class signature. Each parameter is either - recursively cloned if it's a River classes. - deep-copied via copy.deepcopy if not. If the calling object is stochastic (i.e. it accepts a seed parameter) and has not been seeded, then the clone will not be idempotent. Indeed, this method's purpose if simply to return a new instance with the same input parameters.

cond_proba

The underlying data structure used to monitor the input does not allow probability density estimations. Hence, it always returns zero for any given input.

Parameters

  • att_val
  • target_val (Union[bool, str, int])
update

Update statistics of this observer given an attribute value, its target value and the weight of the instance observed.

Parameters

  • att_val
  • target_val (Union[bool, str, int, numbers.Number])
  • sample_weight (float)

References

  1. Domingos, P. and Hulten, G., 2000, August. Mining high-speed data streams. In Proceedings of the sixth ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 71-80). ↩

  2. Pfahringer, B., Holmes, G. and Kirkby, R., 2008, May. Handling numeric attributes in hoeffding trees. In Pacific-Asia Conference on Knowledge Discovery and Data Mining (pp. 296-307). Springer, Berlin, Heidelberg. ↩