Implementation:Scikit learn Scikit learn MulticlassUtils

Knowledge Sources	Scikit_learn Scikit-learn Docs
Domains	Machine Learning, Classification
Last Updated	2026-02-08 15:00 GMT

Overview

Concrete utility module for handling multiclass and multioutput target arrays provided by scikit-learn.

Description

The multiclass module provides utilities for determining and validating the type of target variable in classification tasks. Key functions include unique_labels for extracting ordered unique labels, type_of_target for inferring the target type (binary, multiclass, multilabel-indicator, etc.), and check_classification_targets for validating that targets are suitable for classification.

Usage

Use these utilities when you need to determine the type of classification target, extract unique labels from target arrays, or validate that target data is suitable for a classification estimator.

Code Reference

Source Location

• Repository: scikit-learn
• File: sklearn/utils/multiclass.py

Signature

def unique_labels(*ys):
    ...

def type_of_target(y, input_name=""):
    ...

def check_classification_targets(y):
    ...

def _check_partial_fit_first_call(clf, classes=None):
    ...

def _ovr_decision_function(predictions, confidences, n_classes):
    ...

Import

from sklearn.utils.multiclass import unique_labels, type_of_target, check_classification_targets

I/O Contract

Inputs

Name	Type	Required	Description
ys	array-likes	Yes	One or more label arrays to extract unique labels from
y	array-like	Yes	Target values to determine the type of
input_name	str	No	Name of the input for error messages

Outputs

Name	Type	Description
unique	ndarray	Ordered array of unique labels (for unique_labels)
target_type	str	One of "binary", "multiclass", "multilabel-indicator", "multiclass-multioutput", "continuous", "continuous-multioutput", "unknown"

Usage Examples

Basic Usage

from sklearn.utils.multiclass import unique_labels, type_of_target

# Determine target type
y_binary = [0, 1, 0, 1]
print(type_of_target(y_binary))  # 'binary'

y_multi = [0, 1, 2, 1]
print(type_of_target(y_multi))  # 'multiclass'

# Extract unique labels
labels = unique_labels([3, 5, 5, 5, 7, 7])
print(labels)  # array([3, 5, 7])

Related Pages

• Principle:Scikit_learn_Scikit_learn_Classification_Prediction