bluebell

Basic linear uncertainty estimation with bounding ellipsoids

bluebell is a Python package for estimating the uncertainties of parameters in least-squares problems by computing minimum-volume bounding ellipsoids (MVBEs). Given a set of points at which some has been evaluated, bluebell computes the uncertainties on the points by computing the MVBE around the points, moved by their values of such that, if the problem is exactly linear, all the points lie on the same ellipsoid.

(Source code, png, hires.png, pdf)

bluebell also provides a number of support functions to work with the points and ellipsoids, as well as some plotting functions in bluebell.plot.

You can install bluebell from the Python Package Index (PyPI) with

python -m pip install bluebell

perhaps also with the --user flag, depending how you administer your system. The development version is on GitHub.

API

• bluebell
  • MVBE()
  • select_by_relative_chi2()
  • estimate()
  • propagate()
  • linearize()
  • linear_optimum()
  • uniform_on_unit_sphere()
  • uniform_on_ellipsoid()
  • uniform_in_ellipsoid()
  • vertices()
  • is_in_ellipsoid()
  • is_in_simplex()
  • discard()
  • grid()
• bluebell.plot
  • cov_ellipse()
  • chi2_corner()

Indices and tables

• Index

• Module Index

• Search Page