AI Potluck
Contribute
Back to Gap Map Infrastructure / Classic ML & computer vision

River

online-ml
open source / Overall score: 3.0

Python library for online machine learning, where models learn from one observation at a time on streaming data instead of from a fixed dataset. It covers linear models, trees and forests, naive Bayes, nearest neighbors, clustering, anomaly and drift detection, recommendation, time-series forecasting, bandits and ensembles, with streaming metrics and preprocessing. It came from the merger of creme and scikit-multiflow.

Openness

5 high confidence
5.0
license
BSD-3-Clause(OSI)
source
public(github.com/online-ml/river)
core features withheld
no — community project

BSD-3-Clause and built from the public repository, maintained by its contributors with no company selling an edition.

Adoption

3 high confidence
3.0

Measured on monthly PyPI downloads of the river package.

Capability

3 high confidence
3.0

River is to streaming data what scikit-learn is to fixed datasets, a broad catalog of learners and the tools to evaluate them, and it goes as far as picking among candidate models online. Designing the candidates is still left to the user.

  • https://riverml.xyz/latest/api/overview/ recorded 2026-09-27

    API overview modules include anomaly, bandit, cluster, drift, ensemble, forest, linear_model, model_selection, naive_bayes, neighbors, reco, time_series and tree; "Model selection can be used for tuning the hyperparameters of a model."

Verified 2026-09-27