petsc4py.PETSc.Regressor

class petsc4py.PETSc.Regressor

Bases: Object

Regression solver.

REGRESSOR is described in the PETSc manual.

See also

PetscRegressor

Enumerations

LinearType	Linear regressor type.
Type	REGRESSOR solver type.

Methods Summary

create([comm])	Create a REGRESSOR solver.
destroy()	Destroy the solver.
fit(X, y)	Fit the regression problem.
getLinearCoefficients()	Get a vector of the fitted coefficients from a linear regression model.
getLinearIntercept()	Get the intercept from a linear regression model.
getLinearKSP()	Returns the KSP context used by the linear regressor.
getLinearType()	Return the type of the linear regressor.
getTAO()	Return the underlying TAO object .
getType()	Return the type of the solver.
predict(X, y)	Predict the regression problem.
reset()	Destroy internal data structures of the solver.
setFromOptions()	Configure the solver from the options database.
setLinearFitIntercept(flag)	Set a flag to indicate that the intercept should be calculated.
setLinearType(lineartype)	Set the type of linear regression to be performed.
setLinearUseKSP(flag)	Set a flag to indicate that KSP instead of TAO solvers should be used.
setRegularizerWeight(weight)	Set the weight to be used for the regularizer.
setType(regressor_type)	Set the type of the solver.
setUp()	Set up the internal data structures for using the solver.
view([viewer])	View the solver.

Methods Documentation

create(comm=None)

Create a REGRESSOR solver.

Collective.

Parameters:

comm – MPI communicator, defaults to Sys.getDefaultComm.

Return type:

Self

See also

Sys.getDefaultComm, PetscRegressorCreate

Source code at petsc4py/PETSc/Regressor.pyx:61

destroy()

Destroy the solver.

Collective.

See also

PetscRegressorDestroy

Source code at petsc4py/PETSc/Regressor.pyx:180

Return type:

Self

fit(X, y)

Fit the regression problem.

Collective.

Parameters:

• X (Mat) – The matrix of training data

• y (Vec) – The vector of target values from the training dataset

Return type:

None

See also

PetscRegressorPredict

Source code at petsc4py/PETSc/Regressor.pyx:116

getLinearCoefficients()

Get a vector of the fitted coefficients from a linear regression model.

Not collective.

See also

getLinearIntercept, PetscRegressorLinearGetCoefficients

Source code at petsc4py/PETSc/Regressor.pyx:266

Return type:

Vec

getLinearIntercept()

Get the intercept from a linear regression model.

Not collective.

See also

setLinearFitIntercept, PetscRegressorLinearGetIntercept

Source code at petsc4py/PETSc/Regressor.pyx:280

Return type:

Scalar

getLinearKSP()

Returns the KSP context used by the linear regressor.

Not collective.

See also

PetscRegressorLinearGetKSP

Source code at petsc4py/PETSc/Regressor.pyx:252

Return type:

KSP

getLinearType()

Return the type of the linear regressor.

Not collective.

See also

setLinearType, PetscRegressorLinearGetType

Source code at petsc4py/PETSc/Regressor.pyx:304

Return type:

RegressorLinearType

getTAO()

Return the underlying TAO object .

Not collective.

See also

getLinearKSP, PetscRegressorGetTao

Source code at petsc4py/PETSc/Regressor.pyx:154

Return type:

TAO

getType()

Return the type of the solver.

Not collective.

See also

setType, PetscRegressorGetType

Source code at petsc4py/PETSc/Regressor.pyx:212

Return type:

str

predict(X, y)

Predict the regression problem.

Collective.

Parameters:

• X (Mat) – The matrix of unlabeled observations

• y (Vec) – The vector of predicted labels

Return type:

None

See also

PetscRegressorFit

Source code at petsc4py/PETSc/Regressor.pyx:135

reset()

Destroy internal data structures of the solver.

Collective.

See also

PetscRegressorDestroy

Source code at petsc4py/PETSc/Regressor.pyx:168

Return type:

None

setFromOptions()

Configure the solver from the options database.

Collective.

See also

Working with PETSc options, PetscRegressorSetFromOptions

Source code at petsc4py/PETSc/Regressor.pyx:93

Return type:

None

setLinearFitIntercept(flag)

Set a flag to indicate that the intercept should be calculated.

Logically collective.

See also

PetscRegressorLinearSetFitIntercept

Source code at petsc4py/PETSc/Regressor.pyx:228

Parameters:

flag (bool)

Return type:

None

setLinearType(lineartype)

Set the type of linear regression to be performed.

Logically collective.

See also

getLinearType, PetscRegressorLinearSetType

Source code at petsc4py/PETSc/Regressor.pyx:293

Parameters:

lineartype (RegressorLinearType)

Return type:

None

setLinearUseKSP(flag)

Set a flag to indicate that KSP instead of TAO solvers should be used.

Logically collective.

See also

PetscRegressorLinearSetUseKSP

Source code at petsc4py/PETSc/Regressor.pyx:240

Parameters:

flag (bool)

Return type:

None

setRegularizerWeight(weight)

Set the weight to be used for the regularizer.

Logically collective.

See also

setType, PetscRegressorSetRegularizerWeight

Source code at petsc4py/PETSc/Regressor.pyx:82

Parameters:

weight (float)

Return type:

None

setType(regressor_type)

Set the type of the solver.

Logically collective.

Parameters:

regressor_type (Type | str) – The type of the solver.

Return type:

None

See also

getType, PetscRegressorSetType

Source code at petsc4py/PETSc/Regressor.pyx:193

setUp()

Set up the internal data structures for using the solver.

Collective.

See also

PetscRegressorSetUp

Source code at petsc4py/PETSc/Regressor.pyx:104

Return type:

None

view(viewer=None)

View the solver.

Collective.

Parameters:

viewer (Viewer | None) – A Viewer instance or None for the default viewer.

Return type:

None

See also

PetscRegressorView

Source code at petsc4py/PETSc/Regressor.pyx:42