x3cflux.HopsyModel

class x3cflux.HopsyModel(simulator)

Bases: object

Wrapper to pass ILE sampling problem to hopsy.

Methods

`log_curvature`(x)	Compute the expected Fisher information.
`log_density`(x)	Compute the log likelihood up to a constant.
`log_gradient`(x)	Compute the gradient of the log likelihood.

log_density(x)

Compute the log likelihood up to a constant.

This is equal to

\[-\frac{1}{2} SSR(\boldsymbol \theta)\]

log_gradient(x)

Compute the gradient of the log likelihood.

This is equal to

\[\frac{\partial}{\partial \boldsymbol \theta} \left(-\frac{1}{2} SSR(\boldsymbol \theta)\right)\]

log_curvature(x)

Compute the expected Fisher information.

This is the second moment of the score and is equal to the linearized local Hessian of the SSR.