ML model visualizer

Decision Tree Classifier

Build a tree step by step and see how feature splits create classification regions.

TaskClassificationCriterionGiniExplainabilityHighOutputDecision Rules

Decision Tree Workspace

Decision Space

-2-1012-1.5-0.50.51.5XY
Class AClass BCurrent split

Decision Tree

Class An=80, 40:40
Current split:Root nodeGini gain:0.000Progress:1 / 9 nodesBuilding

Model Metrics

Accuracy50%Tree depth0Nodes1Leaves1Samples80

Current Node

Samples80Gini0.500Best featureFeature 2Threshold-0.226Predicted classClass A

Tree Building

  1. Evaluate every feature
  2. Test candidate thresholds
  3. Choose the split with the highest impurity gain
  4. Partition samples into left and right children
  5. Repeat until a stopping rule is reached
Recursive binary splitsMax depth controls complexityLeaves produce predictions

Split Quality

Parent impurity0.500Left impurity0.062Right impurity0.325Information gain0.277