altar.models.seismic.Static

Module Contents

Classes

class altar.models.seismic.Static.Static(name, locator, **kwds)

Bases: altar.models.bayesian

Static inversion with cuda (d = G theta) Modeled as N patches with dip and slip displacements

parametersets

doc = the set of parameters in the model

parameters

doc = total number of parameters in the model

observations

doc = the number of data samples

patches

norm

default

doc = the norm to use when computing the data log likelihood

case

doc = the directory with the input files

green_file

doc = the name of the file with the Green functions

data_file

doc = the name of the file with the observations

cd_file

doc = the name of the file with the data covariance matrix

output_path

ifs

G

d

Cd

Cd_inv

residuals

normalization = 1

processor = cpu

initialize(self, application)

Initialize the state of the model given a {problem} specification

initializeSample(self, step)

Fill {step.θ} with an initial random sample from my prior distribution.

computePrior(self, step)

Fill {step.prior} with the densities of the samples in {step.theta} in the prior distribution

computeDataLikelihood(self, step)

Fill {step.data} with the densities of the samples in {step.theta} given the available data. This is what is usually referred to as the “forward model”

verify(self, step, mask)

Check whether the samples in {step.theta} are consistent with the model requirements and update the {mask}, a vector with zeroes for valid samples and non-zero for invalid ones

initializeParameterSets(self)

Initialize the parameter set

mountInputDataspace(self, pfs)

Mount the directory with my input files

loadInputs(self)

Load the data in the input files into memory

initializeCovariance(self, samples)

Compute the Cholesky decomposition of the inverse of the data covariance and merge it to data

computeCovarianceInverse(self, cd)

Compute the inverse of the data covariance matrix

computeNormalization(self, observations, cd)

Compute the normalization of the L2 norm

initializeResiduals(self, samples, data)

Prime the matrix that will hold the residuals (G θ - d) for each sample by duplicating the observation vector as many times as there are samples

update(self, annealer)

Model updating at the bottom of each annealing step Output step data

forwardModel(theta, green, data_residuals=None, data_observations=None, batches=None)

Forward model: compute data prediction or data residuals from a set of theta :param theta [in: :type theta [in: samples, parameters :param cuarray] parameters with shape=: :type cuarray] parameters with shape=: samples, parameters :param green [in: :type green [in: observations, parameters :param cuarray] Green’s function with shape =: :type cuarray] Green’s function with shape =: observations, parameters :param batches [in: :param integer: :param optional] number of samples needto be computed <=samples: :param data_observations [in: :param cuarray: :param optional] data observations: :param data_residuals[inout: :type data_residuals[inout: observations, samples :param cuarray: :type cuarray: observations, samples :param optional] data predictions or residuals shape=: :type optional] data predictions or residuals shape=: observations, samples

Returns:

data predictions or residuals if data_observations is provides