Implementation:Online ml River Stream Iter Sklearn

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

Overview

Converts scikit-learn dataset objects (Bunch) into River-compatible data streams.

Description

The iter_sklearn_dataset function enables iteration over scikit-learn datasets, including those loaded via fetch_openml which provides access to thousands of datasets from OpenML. It automatically handles both array and DataFrame formats, extracting feature names when available. This bridges scikit-learn's batch learning interface with River's online learning paradigm.

Usage

Use this when working with scikit-learn's built-in datasets or datasets fetched from OpenML. It's particularly useful for comparing batch and online learning approaches on the same data, or for converting standard benchmark datasets to streaming format.

Code Reference

Source Location

Repository: Online_ml_River
File: river/stream/iter_sklearn.py

Signature

def iter_sklearn_dataset(
    dataset: sklearn.utils.Bunch,
    **kwargs
) -> base.typing.Stream:
    ...

Import

from river import stream

I/O Contract

Parameter	Type	Description
dataset	sklearn.utils.Bunch	A scikit-learn dataset object
**kwargs	dict	Additional arguments passed to iter_array or iter_pandas

Returns:

Type	Description
Iterator[(dict, Any)]	Stream of (features dict, target) tuples

Usage Examples

import pprint
from sklearn import datasets
from river import stream

# Load a scikit-learn dataset
dataset = datasets.load_diabetes()

# Iterate over the dataset
for xi, yi in stream.iter_sklearn_dataset(dataset):
    pprint.pprint(xi)
    print(f"Target: {yi}")
    break
# Output:
# {'age': 0.038075906433423026,
#  'bmi': 0.061696206518683294,
#  'bp': 0.0218723855140367,
#  's1': -0.04422349842444599,
#  's2': -0.03482076283769895,
#  's3': -0.04340084565202491,
#  's4': -0.002592261998183278,
#  's5': 0.019907486170462722,
#  's6': -0.01764612515980379,
#  'sex': 0.05068011873981862}
# Target: 151.0

# Using with other sklearn datasets
iris = datasets.load_iris()
print("\nIris dataset:")
for x, y in stream.iter_sklearn_dataset(iris):
    print(f"Features: {list(x.keys())}")
    print(f"Target: {y}")
    break

# With OpenML datasets (requires internet)
# from sklearn.datasets import fetch_openml
# credit_g = fetch_openml('credit-g', version=1, parser='auto')
# for x, y in stream.iter_sklearn_dataset(credit_g):
#     print(x, y)
#     break

# You can pass additional kwargs like shuffle
wine = datasets.load_wine()
shuffled = stream.iter_sklearn_dataset(wine, shuffle=True, seed=42)

print("\nFirst 3 shuffled samples:")
for i, (x, y) in enumerate(shuffled):
    if i >= 3:
        break
    print(f"Sample {i}: target={y}, features={list(x.keys())}")

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment