Machine Learning Uncertainty Estimation Toolbox

David A. Sheen

National Institute of Standards and Technology

Last Revision Jun 10, 2019

Download this software from GitHub

Welcome to the home page for the Machine Learning Uncertainty Toolbox. This software is a Python package that calculates uncertainties in machine learning models using bootstrapping and residual bootstrapping. It is intended to interface with scikit-learn but any Python package that uses a similar interface should work.

Contents

Contents:

• Module Reference
  • Function Summary
  • Function documentation
• Example linear regression (1st-order polynomial)
  • Imports
  • Data and Metadata
  • Regression
  • Uncertainty Calculation
• Example linear regression (2nd-order polynomial)
  • Imports
  • Data and Metadata
  • Regression
  • Uncertainty Calculation
• Uncertainty in QSAR modeling
• Uncertainty in biomarker identification
  • Imports and local function definitions
  • Consensus analysis
  • Uncertainty analysis on unsupervised classification
  • Uncertainty analysis on supervised learning

Legal

This software is subject to the NIST Software License (revised as of August 2018).

Contact

David Sheen

• Email
• NIST Staff Page
• GitHub profile

Links

NIST GitHub Organization

NIST MetQual program

NIST Chemical Informatics Research Group

NIST home page

Indices and tables

• Index
• Module Index
• Search Page