bayesquad

probnum.quad.bayesquad(fun, fun0, bounds, nevals=None, type='vanilla', **kwargs)[source]

One-dimensional Bayesian quadrature.

Parameters:
  • fun (function) – Function to be integrated.
  • fun0 (RandomProcess) – Stochastic process modelling the function to be integrated.
  • bounds (ndarray, shape=(2,)) – Lower and upper limit of integration.
  • nevals (int) – Number of function evaluations.
  • type (str) –

    Type of Bayesian quadrature to use. The available options are

    vanilla vanilla
    WSABI wsabi
  • kwargs (optional) – Keyword arguments passed on.
Returns:

  • F (RandomVariable) – The integral of func from a to b.
  • fun0 (RandomProcess) – Stochastic process modelling the function to be integrated after neval observations.
  • info (dict) – Information on the performance of the method.

References

See also

nbayesquad()
N-dimensional Bayesian quadrature.