Caesar AI Atlas

Threshold

Also known as: Threshold (for Decision Trees) Β· for Decision Trees Β· Threshold for Decision Trees

Caesar AI Atlas Definition

A threshold for decision trees is the comparison value used in a condition on a numerical feature. It determines how examples are split at a node, such as whether a feature value is above or below the chosen cutoff.

Other Definitions

Threshold [for Decision Trees] Source

In an axis-aligned condition, the value that a feature is being compared against. For example, 75 is the threshold value in the following condition: [!NOTE] This form of the term threshold is different than classification threshold. See Exact splitter for binary classification with numerical features in the Decision Forests course for more information.

Related Terms