← All posts

What Is the Split Rule (Discrimination Rule) in Random Forest? Gini vs Extra Trees Explained

Clinical Epidemiology ResearchUniqcret doctor knowledgesMethodology and Research DesignMethodologic [Methodology]
On this page

What is the Split Rule?

At each node in a decision tree, the algorithm must decide:

“Where should I split this feature to best separate the outcome?”

This decision is governed by the split rule (criterion).

In Random Forest, the most common split rules are:

The key difference lies in how the split threshold is chosen.


The Core Difference: How a Split Point is Chosen

Consider a single feature:

Standard Random Forest (Gini impurity)

Process:

Candidate split pointLeft groupRight groupResult
Between 22–35[22][35, 41, 55, 63, 70, 78]Evaluate impurity
Between 35–41[22, 35][41, 55, 63, 70, 78]Evaluate impurity
Between 41–55[22, 35, 41][55, 63, 70, 78]Evaluate impurity
Between 55–63[22, 35, 41, 55][63, 70, 78]Evaluate impurity

The algorithm:

Interpretation:

Behavior
Exhaustive search
Always selects the optimal split
Deterministic given the data

Extra Trees (Extremely Randomized Trees)

Process:

StepAction
1Randomly generate one split point within the feature range
2Apply that split directly
3Do not compare with other candidates

Example:

Interpretation:

Behavior
No search for the best split
Uses one random threshold
Stochastic (random) decision

Visual Analogy

Target analogy

MethodStrategy
Gini (Standard RF)Tests many positions and selects the best
Extra TreesPicks one random position

What Happens Across the Forest

Standard Random Forest

PropertyBehavior
Feature selectionRandom (controlled by mtry)
Data samplingBootstrap
Split selectionOptimal (deterministic)

Result:

Extra Trees

PropertyBehavior
Feature selectionRandom
Data samplingBootstrap
Split selectionRandom threshold

Result:


Bias–Variance Trade-off

MethodBiasVarianceExplanation
Single decision treeLowHighOverfits data
Standard Random ForestModerateModerateBalanced
Extra TreesSlightly higherLowerMore randomness reduces variance

Interpretation


Effect on Individual Trees

Standard Random Forest

Tree 1Tree 2
Age < 48Age < 48
SBP < 120SBP < 125

Pattern:

Extra Trees

Tree 1Tree 2
Age < 52Age < 37
SBP < 108SBP < 135

Pattern:


When Each Approach Performs Better

ScenarioStandard RF (Gini)Extra Trees
Small datasetBetterAcceptable
Large datasetGoodBetter
Few predictorsBetterAcceptable
Many predictorsGoodBetter
Training speedSlowerFaster

Practical Impact on Model Performance

In most real-world clinical prediction settings:

ModelTypical AUROC
Standard Random Forest~0.78
Extra Trees~0.77

Difference:

From a prediction modeling perspective:

not by the split rule itself.


Interpretation for Clinical Prediction Models

From a methodological standpoint:


Practical Recommendation


Key Takeaways

0
Message for International and Thai ReadersUnderstanding My Medical Context in ThailandRead more →Message for International and Thai ReadersUnderstanding My Broader Content Beyond MedicineRead more →

Comments

No comments yet. Be the first to share your thoughts.

Sign in to comment