cma.sampler.GaussSampler

class documentation

class GaussSampler(StatisticalModelSamplerWithZeroMeanBaseClass):

Known subclasses: cma.sampler.GaussDiagonalSampler, cma.sampler.GaussFullSampler, cma.sampler.GaussStandardConstant

View In Hierarchy

Undocumented

Method __init__ declarative init, doesn't need to be executed
Method set_H set Hessian w.r.t. which to compute the eigen spectrum.
Method set_H_by_f set Hessian from f at x0.
Instance Variable dimension Undocumented
Property chin approximation of the expected length when isotropic with variance 1.
Property corr_condition condition number of the correlation matrix
Property eigenspectrum return eigen spectrum w.r.t. H like sqrt(H) C sqrt(H)
Instance Variable _left Undocumented
Instance Variable _right Undocumented

Inherited from StatisticalModelSamplerWithZeroMeanBaseClass:

Method __imul__ Undocumented
Method inverse_hessian_scalar_correction return scalar correction alpha such that X and f fit to f(x) = (x-mean) (alpha * C)**-1 (x-mean)
Method norm return Mahalanobis norm of x w.r.t. the statistical model
Method parameters return dict with (default) parameters, e.g., c1 and cmu.
Method sample return list of i.i.d. samples.
Method to_linear_transformation return associated linear transformation
Method to_linear_transformation_inverse return inverse of associated linear transformation
Method transform transform x as implied from the distribution parameters
Method transform_inverse Undocumented
Method update vectors is a list of samples, weights a corrsponding list of learning rates
Property condition_number Undocumented
Property covariance_matrix Undocumented
Property variances vector of coordinate-wise (marginal) variances
Instance Variable _lam Undocumented
Instance Variable _mueff Undocumented
Instance Variable _parameters Undocumented
def __init__(self):
overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.__init__
overridden in cma.sampler.GaussDiagonalSampler, cma.sampler.GaussFullSampler, cma.sampler.GaussStandardConstant

declarative init, doesn't need to be executed

def set_H(self, H):

set Hessian w.r.t. which to compute the eigen spectrum.

def set_H_by_f(self, f, x0, eps=None):

set Hessian from f at x0.

>>> import numpy as np, cma
>>> es = cma.CMAEvolutionStrategy(3 * [1], 1, {'verbose':-9})
>>> es.sm.set_H_by_f(cma.ff.elli, 3 * [0])  # Hessian of cma.ff.elli

Now the eigen spectrum of H^1/2 C H^1/2 where H is the Hessian of cma.ff.elli is given by the spectrum property.

dimension: int =
overridden in cma.sampler.GaussDiagonalSampler, cma.sampler.GaussFullSampler, cma.sampler.GaussStandardConstant

Undocumented

@property
chin =

approximation of the expected length when isotropic with variance 1.

The exact value could be computed by:

from scipy.special import gamma
return 2**0.5 * gamma((self.dimension+1) / 2) / gamma(self.dimension / 2)

The approximation obeys chin < chin_hat < (1 + 5e-5) * chin.

@property
corr_condition =
overridden in cma.sampler.GaussFullSampler

condition number of the correlation matrix

@property
eigenspectrum =

return eigen spectrum w.r.t. H like sqrt(H) C sqrt(H)

_left =

Undocumented

_right =

Undocumented