Home | Trees | Indices | Help |
---|
|
The :mod:`parsimony.functions.penalties` module contains the penalties used to constrain the loss functions. These represent mathematical functions and should thus have properties used by the corresponding algorithms. These properties are defined in :mod:`parsimony.functions.properties`.
Penalties should be stateless. Penalties may be shared and copied and should therefore not hold anything that cannot be recomputed the next time it is called.
Created on Mon Apr 22 10:54:29 2013
Copyright (c) 2013-2014, CEA/DSV/I2BM/Neurospin. All rights reserved.
Author: Tommy Löfstedt, Vincent Guillemot, Edouard Duchesnay and Fouad Hadj-Selem
License: BSD 3-clause.
|
|||
ZeroFunction | |||
L1 The proximal operator of the L1 function with a penalty formulation |
|||
L0 The proximal operator of the "pseudo" L0 function |
|||
LInf The proximal operator of the L-infinity function |
|||
L2 The proximal operator of the L2 function with a penalty formulation |
|||
L2Squared The proximal operator of the squared L2 function with a penalty formulation |
|||
L1L2Squared The proximal operator of the L1 function with an L2 constraint. |
|||
QuadraticConstraint The proximal operator of the quadratic function |
|||
RGCCAConstraint Represents the quadratic function |
|||
RidgeSquaredError Represents a ridge squared error penalty, i.e. |
|||
LinearVariableConstraint Represents a linear constraint |
|||
SufficientDescentCondition |
|
|||
__package__ =
|
Home | Trees | Indices | Help |
---|
Generated by Epydoc 3.0.1 on Mon Apr 6 23:52:10 2015 | http://epydoc.sourceforge.net |