fitbenchmarking.hessian.scipy_hessian module

Module which calculates SciPy finite difference approximations

class fitbenchmarking.hessian.scipy_hessian.Scipy(problem, jacobian)

Bases: Hessian

Implements SciPy finite difference approximations to the derivative

INCOMPATIBLE_PROBLEMS = {'cs': ['mantid']}

eval(params, **kwargs)

Evaluates Hessian of problem.eval_model, returning the value

abla^2_p f(x, p)

param params:

The parameter values to find the Hessian at

type params:

list

return:

Approximation of the Hessian

rtype:

3D numpy array