Implementation:Online ml River Tree Nodes Branch

Knowledge Sources	Online_ml_River
Domains	Online_Learning, Decision_Trees
Last Updated	2026-02-08 16:00 GMT

Overview

Branch node implementations for decision trees supporting binary and multiway splits on both numeric and nominal features.

Description

This module provides the branch (split) node implementations for decision trees. DTBranch is the base class, with four concrete implementations: NumericBinaryBranch (threshold-based splits on numerical features), NominalBinaryBranch (equality splits on categorical features), NumericMultiwayBranch (interval-based multiway splits using radius), and NominalMultiwayBranch (multiway splits on categorical values). Each branch type determines how to route samples to children and provides methods for finding the most common path.

Usage

Use these branch nodes when building decision tree structures. Binary branches are standard for most trees, while multiway branches can reduce tree depth but increase branching factor.

Code Reference

Source Location

• Repository: Online_ml_River
• File: river/tree/nodes/branch.py

Signature

class DTBranch(Branch):
    def __init__(self, stats, *children, **attributes):
        ...
    def branch_no(self, x):
        pass
    def next(self, x):
        ...
    def max_branches(self):
        pass
    def repr_branch(self, index: int, shorten=False):
        pass

class NumericBinaryBranch(DTBranch):
    def __init__(self, stats, feature, threshold, depth, left, right, **attributes):
        ...
    def branch_no(self, x):
        return 0 if x[self.feature] <= self.threshold else 1
    def most_common_path(self):
        ...

class NominalBinaryBranch(DTBranch):
    def __init__(self, stats, feature, value, depth, left, right, **attributes):
        ...
    def branch_no(self, x):
        return 0 if x[self.feature] == self.value else 1

class NumericMultiwayBranch(DTBranch):
    def __init__(self, stats, feature, radius_and_slots, depth, *children, **attributes):
        ...
    def add_child(self, feature_val, child):
        ...

class NominalMultiwayBranch(DTBranch):
    def __init__(self, stats, feature, feature_values, depth, *children, **attributes):
        ...
    def add_child(self, feature_val, child):
        ...

Import

from river.tree.nodes.branch import NumericBinaryBranch
from river.tree.nodes.branch import NominalBinaryBranch
from river.tree.nodes.branch import NumericMultiwayBranch
from river.tree.nodes.branch import NominalMultiwayBranch

I/O Contract

Input	Type	Description
stats	dict/Var	Split node statistics
feature	str/int	Feature name or index
threshold/value	float/any	Split threshold or categorical value
depth	int	Node depth
children	tuple	Child nodes

Output	Type	Description
branch_no	int	Child branch index for routing
next	Node	Next child node for given sample
repr_branch	str	String representation of branch test

Usage Examples

from river.tree.nodes.branch import NumericBinaryBranch, NominalBinaryBranch
from river.tree.nodes.htc_nodes import LeafMajorityClass
from river.tree.splitter import GaussianSplitter

# Create leaves
left_leaf = LeafMajorityClass({'A': 10}, 2, GaussianSplitter())
right_leaf = LeafMajorityClass({'B': 8}, 2, GaussianSplitter())

# Create numeric binary branch
branch = NumericBinaryBranch(
    stats={'A': 10, 'B': 8},
    feature='age',
    threshold=30.0,
    depth=1,
    left=left_leaf,
    right=right_leaf
)

# Route sample
x = {'age': 25, 'income': 50000}
branch_idx = branch.branch_no(x)  # Returns 0 (left)
next_node = branch.next(x)  # Returns left_leaf

# Create nominal binary branch
nom_branch = NominalBinaryBranch(
    stats={'A': 5, 'B': 5},
    feature='color',
    value='red',
    depth=1,
    left=left_leaf,
    right=right_leaf
)

x2 = {'color': 'red'}
branch_idx = nom_branch.branch_no(x2)  # Returns 0 (equal)

Related Pages

• Environment:Online_ml_River_Python_Runtime_Environment