This Swift library builds machine learning predictions using linear regression.
Open source
24 alternativesMCGrad
MCGrad calibrates machine learning predictions across the full dataset and across segments defined by input features.
About MCGrad
MCGrad is an open-source Python tool for multicalibration of machine learning model predictions. Traditional calibration methods such as Isotonic Regression and Platt Scaling ensure calibration only across all data. MCGrad also aims to calibrate predictions across segments defined by input features, such as country or content type, without requiring the user to specify protected groups manually. Installation requires Python 3.10 or later and can be done with pip or conda. The README states that MCGrad has been deployed at Meta on hundreds of production models, and points to a research paper for detailed experimental results.
Key features
- Calibrates predictions across all data and across segments defined by input features
- Identifies calibration gaps in segments without manual specification of protected groups
- Fit and predict methods that work on pandas DataFrames with named columns
- Supports categorical feature columns such as country or content type
- Validation-based early stopping for likelihood-improving updates
- Interactive notebooks, including one on calibrating LLM confidence scores
Best for: Machine learning practitioners working in Python who need predicted probabilities to be well calibrated across subgroups of their data, not only overall.
Good for
- Open source, with the code hosted in a public GitHub repository
- Takes input features directly rather than predefined segment definitions
- Documentation includes a quick start, methodology overview and API reference
Watch out for
- Requires Python 3.10 or later
MCGrad at a glance
- Category
- AI & Machine Learning › AI Data & Analytics
- Type
- Open-source project
- Pricing
- Not confirmed. Check the site
- Open source
- Yes · Source code
- Language
- English
- Last checked
- Oct 6, 2026 · Online when checked
- Listed
- Oct 5, 2026
MCGrad alternatives
Tools and sites you could use instead, ranked by how closely they match.
Python package which helps to debug machine learning classifiers and explain their predictions.
Open source
24 alternatives
LIME provides explanations for predictions made by machine learning classifiers.
Open source
24 alternatives
PredictionBuilder builds machine learning predictions using linear regression.
Open source
24 alternatives
19 Questions assigns attributes to objects using machine learning and Bayesian inference, according to its project description.
Open source
24 alternatives
oneDAL is a C++ and DPC++ library for accelerated machine learning routines on tabular data, governed by the UXL Foundation.
Open source
24 alternativesPopular in AI Data & Analytics
Well-known listings from the same topic.
Dotgetdot.aiForvioforvio.comFluidAI Medicalfluidai.mdApache Zeppelinzeppelin.apache.orgSola Securitysola.securityQuadraticquadratichq.com
MCGrad FAQ
What does calibration mean in this context?
Calibration means predicted probabilities match observed outcomes on average. MCGrad applies this across all data and across segments defined by features.
How can MCGrad be installed?
The stable release can be installed with pip install mcgrad, or with conda install mcgrad or mamba install mcgrad. Python 3.10 or later is required.
Where can questions or bugs be reported?
Questions and bugs can be reported by opening an issue on GitHub Issues.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.