rxmc.proposal.LogspaceNormalProposalDistribution#

class rxmc.proposal.LogspaceNormalProposalDistribution(scale: float)[source]#

Bases: ProposalDistribution

Normal proposal operating in log space, for strictly positive parameters.

Proposes exp(log(x) + eps) where eps ~ Normal(0, scale), which preserves positivity while allowing multiplicative jumps of arbitrary size.

Parameters:

scale (float) – Standard deviation of the normal perturbation in log space.

__init__(scale: float)[source]#

Methods

__init__(scale)