atm.method module

Classes

Categorical(name, type, values)

HyperParameter

HyperPartition(categoricals, constants, tunables)

Class which holds the hyperparameter settings that define a hyperpartition.

List(name, type, list_length, element)

Method(method)

This class is initialized with the name of a json configuration file.

Numeric(name, type, range)

class atm.method.Categorical(name, type, values)[source]

Bases: atm.method.HyperParameter

Methods

as_tunable()

Attributes

is_categorical

is_constant

as_tunable()[source]
property is_categorical
property is_constant
class atm.method.HyperParameter[source]

Bases: object

Attributes

is_categorical

is_constant

property is_categorical
property is_constant
class atm.method.HyperPartition(categoricals, constants, tunables)[source]

Bases: object

Class which holds the hyperparameter settings that define a hyperpartition.

class atm.method.List(name, type, list_length, element)[source]

Bases: atm.method.HyperParameter

Methods

get_elements()

Attributes

is_categorical

get_elements()[source]
property is_categorical
class atm.method.Method(method)[source]

Bases: object

This class is initialized with the name of a json configuration file. The config contains information about a classification method and the hyperparameter arguments it needs to run. Its main purpose is to generate hyperpartitions (possible combinations of categorical hyperparameters).

Methods

get_hyperpartitions()

Traverse the CPT and enumerate all possible hyperpartitions of categorical parameters for this method.

get_hyperpartitions()[source]

Traverse the CPT and enumerate all possible hyperpartitions of categorical parameters for this method.

class atm.method.Numeric(name, type, range)[source]

Bases: atm.method.HyperParameter

Methods

as_tunable()

Attributes

is_constant

as_tunable()[source]
property is_constant