EN
Interactive Machine Learning Playground

Train a Real Decision Treeon the Iris Flower Dataset

Use the original Fisher Iris measurements. Watch the tree choose thresholds, reduce impurity, create branches, and classify Setosa, Versicolor, and Virginica samples.

150real flower samples
4numeric measurements in centimeters
3Iris species
Decision TreeTrain with Gini impurity or entropy

Decision Regions

Petal length vs. petal width

The colored regions show a two-feature slice through the four-feature model; other measurements are fixed at their dataset means. Points show all 150 observations. Test accuracy is a learning diagnostic, not a final model-selection benchmark.

Click Train Decision Tree to compute the first split from the real Iris measurements.

Trained Tree

waiting for training
No tree yetChoose a depth and click Train Decision Tree.

What Happens During Training

Gini impurity

Classify a New Flower

follow the path through the tree
Train the tree first, then adjust the measurements to classify a new sample.

Original Iris Records

150 samples