# stanfordmlgroup/ngboost

Natural Gradient Boosting for Probabilistic Prediction

Repository: https://github.com/stanfordmlgroup/ngboost
Canonical: https://ross.abutalabs.com/products/ngboost
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Topics: machine-learning, gradient-boosting, natural-gradients, uncertainty-estimation, ngboost, python
Last push: 2026-07-01T01:46:15+00:00

## Health v2 (maintenance only)
Score: 88/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 90, release rhythm 78, longevity 100
- inputs: {"age_days": 2995, "days_push": 64, "days_rel": 68, "gap_med": 45, "n_releases_24m": 10}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1887, forks 253 (observed 2026-08-28T04:05:49.201289+00:00)

## What it is
NGBoost is a Python library implementing Natural Gradient Boosting for probabilistic prediction, built on top of scikit-learn. It outputs full predictive distributions rather than point estimates, with modular choice of base learner, distribution, and scoring rule.

## Use cases
- predict a full probability distribution instead of a point estimate with gradient boosting
- estimate uncertainty in regression predictions
- fit a gradient boosting model that outputs mean and variance
- probabilistic regression on tabular data in Python
- compare boosting models using negative log likelihood
- get calibrated prediction intervals from tree ensembles

## When to choose
- you need uncertainty quantification from gradient boosting on tabular data
- you want a scikit-learn-compatible probabilistic regressor
- you need to model output distributions like Normal or Poisson with boosting

## When to avoid
- you only need fast point predictions where XGBoost or LightGBM are more optimized
- you need classification with modern GPU-accelerated boosting
- you require very large-scale training with heavy ecosystem support

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, data-science
- domain: machine-learning, data-science
- platform: python
- tags: gradient-boosting, uncertainty-estimation, probabilistic-prediction, natural-gradients, scikit-learn, regression

## Member repositories
- stanfordmlgroup/ngboost (main) score 88

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:49.201289+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:13:33.448040+00:00, confidence not recorded.
  - readme: https://github.com/stanfordmlgroup/ngboost (fetched 2026-08-28T04:05:49.201289+00:00, sha 5aba3571658d)
  - registry_pypi: https://pypi.org/pypi/ngboost/json (fetched 2026-08-29T10:52:32.352249+00:00, sha 0938e5551377)
- Data as of 2026-08-30T08:39:29.467469+00:00.