smica.component.SourceND Class Reference

SourceND. More...

Inheritance diagram for smica.component.SourceND:

Public Member Functions
def	nspec
def	get_dim
def	is_astro
def	uvech
def	vech
def	set_theta
def	get_theta
def	get_thetascore
def	get_thetaroot
def	get_thetaroot_local
def	get_thetafim
def	close_form
Public Member Functions inherited from smica.component.Source
def	is_astro
def	set_dim
def	__init__
def	set_mixmat
def	fix_mixmat
def	set_powspec
def	fix_powspec
def	mixmat
def	powspec
def	autopowspec
def	covariance
def	score_mixmat
def	score_powspec
def	plot_mixmat
def	plot_powspec
Public Member Functions inherited from smica.component.Component
def	plot_power
def	plot_em
def	set_gaussian_prior
def	get_penalty
def	get_score_penalty
def	get_fim_penalty

Properties
	nspec = property(nspec)

Additional Inherited Members
Public Attributes inherited from smica.component.Source
	dim
	ndet
	nbin
	name

Detailed Description

SourceND.

---

A multi dimensional source component.

The mixing matrix is (ndet, dim) and auto and cross power spectra
are described by an array of shape (dim, dim, nbin).  Only the
lower triangular part of each matrices is considered as free
parameters.

Member Function Documentation

def smica.component.SourceND.close_form	(		self,
			stats,
			N
	)

Update local parameters (ie power spectra array) using a close form algorithm.

Parameters
----------
stats : array-like, shape (ndet, ndet, nbin).
Second order statistics ie empirical covariance matrices of
the observations.
N : array-like, shape (ndet, ndet, nbin).
Covariance matrices of the model deprived from the
contribution of the component.

def smica.component.SourceND.get_dim

(

self)

Return the number of free parameters of the component.

def smica.component.SourceND.get_theta	(		self,
			mixmat = None,
			powspec = None
	)

Return component parameters as a vector.

Parameters
----------
mixmat : array-like, shape (ndet, dim).
If None, values are read from inner mixing matrix.
powspec : array-like, shape (dim,dim,nbin).
If None, values are read from inner power spectra.

Returns
----------
array-like, shape (N, 1), where N is the number of free parameters.

def smica.component.SourceND.get_thetafim	(		self,
			iRr,
			w
	)

Return Fisher Information Matrix.

Parameters
----------
iRr : array-like, shape (ndet, ndet, nbin).
iRr is the square root of the inverse of the covariance of the
model, which is constant for all components.
w : array-like, shape (1, nbin). w is the number of modes for each bin.

Returns
----------
array-like, shape (N, N), where N is the number of free parameters.

def smica.component.SourceND.get_thetaroot	(		self,
			iRr,
			bin = 0
	)

Return root contribution to the Fisher Information Matrix for a given bin.

Parameters
----------
iRr : array-like, shape (ndet, ndet).
iRr is the square root of the inverse of the covariance of the
model, which is constant for all components.
bin : int, bin number

Returns
----------
array-like, shape (mxm, N), where N is the number of free parameters.

def smica.component.SourceND.get_thetaroot_local	(		self,
			iRr,
			Bq,
			bin = 0
	)

Return root contribution to the Fisher Information Matrix of
local parameters for a given bin.

Parameters
----------
iRr : array-like, shape (ndet, ndet).
  iRr is the square root of the inverse of the covariance of
  the model, which is constant for all components.

Bq : array-like, shape (ndet, ndet).
  Bq is the beam matrix at bin q

q : int
  The bin number

Returns
-------
  array-like, shape (ndet * ndet, N), where N is the number of
  local free parameters and ndet the number of detectors.

def smica.component.SourceND.get_thetascore	(		self,
			D
	)

Return likelihood first derivative values wrt component
parameters as a vector.

Parameters
----------
D : array-like, shape (ndet, ndet, nbin).
D is the first derivative of the likelihood wrt the covariance
of the model, which is constant for all components.

Returns
----------
array-like, shape (N, 1), where N is the number of free parameters.

def smica.component.SourceND.nspec

(

self)

Return the number of auto and cross spectra.

def smica.component.SourceND.set_theta	(		self,
			theta
	)

Set new values for component parameters.

Parameters
----------
theta : array-like, shape (N, 1), where N is the number of free parameters.

def smica.component.SourceND.uvech	(		self,
			V
	)

Return the matrix of auto and cross spectra from stacking vector.

Convention used is : first vector values (dim) are auto
spectra, then cross spectra (1-2, 1-3, ...).  This convention
is saved into a global variable named 'Midx' in order to save
computational time.

Parameters
----------
V : array-like, shape (dim*(dim+1)/2, 1).
Vector to be convert into a symmetric matrix

Returns
-------
array-like, shape (dim, dim). A symmetric matrix.

def smica.component.SourceND.vech	(		self,
			M
	)

Return the stacking vector of auto and cross spectra from matrix.

Convention used is : first vector values (dim) are auto
spectra, then cross spectra (1-2, 1-3, ...).  This convention
is saved into a global variable named 'Vidx' in order to save
computational time.

Parameters
----------
V : array-like, shape (dim, dim).
A symmetric matrix to be convert into a vector.

Returns
-------
array-like, shape (dim*(dim+1)/2, 1).

---

The documentation for this class was generated from the following file:

• smica/component.py