secretflow.ml.linear.ss_glm.core#

Classes:

LinkType(value)

An enumeration.

DistributionType(value)

An enumeration.

Linker()

Distribution(s)

Functions:

get_link(t)

get_dist(t, scale[, tweedie_power])

class secretflow.ml.linear.ss_glm.core.LinkType(value)[源代码]#

基类：Enum

An enumeration.

Attributes:

Logit

Log

Reciprocal

Indentity

Logit = 'Logit'#
Log = 'Log'#
Reciprocal = 'Reciprocal'#
Indentity = 'Indentity'#
class secretflow.ml.linear.ss_glm.core.DistributionType(value)[源代码]#

基类：Enum

An enumeration.

Attributes:

Bernoulli

Poisson

Gamma

Tweedie

Bernoulli = 'Bernoulli'#
Poisson = 'Poisson'#
Gamma = 'Gamma'#
Tweedie = 'Tweedie'#
secretflow.ml.linear.ss_glm.core.get_dist(t: Union[DistributionType, str], scale: float, tweedie_power: float = 1) → Distribution[源代码]#
class secretflow.ml.linear.ss_glm.core.Linker[源代码]#

基类：ABC

Methods:

link(mu)

response(eta)

response_derivative(mu)

link_derivative(mu)

abstract response(eta: ndarray) → ndarray[源代码]#
abstract response_derivative(mu: ndarray) → ndarray[源代码]#
class secretflow.ml.linear.ss_glm.core.Distribution(s: float)[源代码]#

基类：ABC

Methods:

__init__(s)

scale()

variance(mu)

starting_mu(labels)

deviance(preds, labels[, weights])

__init__(s: float) → None[源代码]#
scale() → float[源代码]#
abstract variance(mu: ndarray) → ndarray[源代码]#
abstract starting_mu(labels: ndarray) → ndarray[源代码]#
abstract deviance(preds: ndarray, labels: ndarray, weights: Optional[ndarray] = None) → ndarray[源代码]#

secretflow.ml.linear.ss_glm.core.distribution#

Classes:

Distribution(s)

DistributionBernoulli(s)

DistributionPoisson(s)

DistributionGamma(s)

DistributionTweedie(s, p)

DistributionType(value)

An enumeration.

Functions:

get_dist(t, scale[, tweedie_power])

class secretflow.ml.linear.ss_glm.core.distribution.Distribution(s: float)[源代码]#

基类：ABC

Methods:

__init__(s)

scale()

variance(mu)

starting_mu(labels)

deviance(preds, labels[, weights])

__init__(s: float) → None[源代码]#
scale() → float[源代码]#
abstract variance(mu: ndarray) → ndarray[源代码]#
abstract starting_mu(labels: ndarray) → ndarray[源代码]#
abstract deviance(preds: ndarray, labels: ndarray, weights: Optional[ndarray] = None) → ndarray[源代码]#
class secretflow.ml.linear.ss_glm.core.distribution.DistributionBernoulli(s: float)[源代码]#

基类：Distribution

Methods:

variance(mu)

starting_mu(labels)

deviance(preds, labels[, weights])

variance(mu: ndarray) → ndarray[源代码]#
starting_mu(labels: ndarray) → ndarray[源代码]#
deviance(preds: ndarray, labels: ndarray, weights: Optional[ndarray] = None) → ndarray[源代码]#
class secretflow.ml.linear.ss_glm.core.distribution.DistributionPoisson(s: float)[源代码]#

基类：Distribution

Methods:

variance(mu)

starting_mu(labels)

deviance(preds, labels[, weights])

variance(mu: ndarray) → ndarray[源代码]#
starting_mu(labels: ndarray) → ndarray[源代码]#
deviance(preds: ndarray, labels: ndarray, weights: Optional[ndarray] = None) → ndarray[源代码]#
class secretflow.ml.linear.ss_glm.core.distribution.DistributionGamma(s: float)[源代码]#

基类：Distribution

Methods:

variance(mu)

starting_mu(labels)

deviance(preds, labels[, weights])

variance(mu: ndarray) → ndarray[源代码]#
starting_mu(labels: ndarray) → ndarray[源代码]#
deviance(preds: ndarray, labels: ndarray, weights: Optional[ndarray] = None) → ndarray[源代码]#
class secretflow.ml.linear.ss_glm.core.distribution.DistributionTweedie(s: float, p: float)[源代码]#

基类：Distribution

Methods:

__init__(s, p)

variance(mu)

starting_mu(labels)

deviance(preds, labels[, weights])

__init__(s: float, p: float)[源代码]#
variance(mu: ndarray) → ndarray[源代码]#
starting_mu(labels: ndarray) → ndarray[源代码]#
deviance(preds: ndarray, labels: ndarray, weights: Optional[ndarray] = None) → ndarray[源代码]#
class secretflow.ml.linear.ss_glm.core.distribution.DistributionType(value)[源代码]#

基类：Enum

An enumeration.

Attributes:

Bernoulli

Poisson

Gamma

Tweedie

Bernoulli = 'Bernoulli'#
Poisson = 'Poisson'#
Gamma = 'Gamma'#
Tweedie = 'Tweedie'#
secretflow.ml.linear.ss_glm.core.distribution.get_dist(t: Union[DistributionType, str], scale: float, tweedie_power: float = 1) → Distribution[源代码]#