grnet.models package

Module contents

class grnet.models.BinPC(data: DataFrame, n: int | None = None, random_state: int = 0)

Bases: Estimator

pgmpy wrapper class for PC algorithm with DOR-based binarization

Methods

__init__(: self, data: pandas.DataFrame, n: int, random_state: int
) -> None:: initialize attributes
estimate(: self, variant: str, ci_test: str, max_cond_vars: int, return_type: str, significance_level: float, n_jobs: int, show_progress: bool
) -> None:: estimate network and save edges (a list of tuples) as self.edges by PC algorithm actual implementation of PC algorithm is a wrapper of pgmpy.estimators.PC default values are altered from the original codes for some arguments to adjust for GRNs
get_matrix(: self
) -> pandas.core.frame.DataFrame:: export network information as DxD matrix of 0 or 1 elements

Attributes

data: pandas.core.frame.DataFrame: input data or resampled data (data will be resampled if n is specified in self.__init__)
model: pgmpy.estimators.PC.PC: model information (for debugging)
edges: List[tuple]: information of edges are saved as a list of tuples after self.estimate was run

References

pgmpy.estimators.PC: https://pgmpy.org/structure_estimator/pc.html?highlight=pc

estimate(variant: str = 'stable', ci_test: str = 'chi_square', max_cond_vars: int | None = None, return_type: str = 'dag', significance_level: float = 0.01, n_jobs: int = -1, show_progress: bool = False) → None

Parameters

variant: str, default: “stable”, ci_test: str, default: “chi_square”, max_cond_vars: int, default: None, return_type: str, default: “dag”, significance_level: float, default: 0.01, n_jobs: int, default: -1, show_progress: bool, default: False

Returns

None

References

pgmpy.estimators.PC: https://pgmpy.org/structure_estimator/pc.html?highlight=pc
grnet.abstract.Estimator: https://grnet.readthedocs.io/en/latest/grnet.abstract.html#grnet.abstract.Estimator

get_matrix() → DataFrame

Parameters

None

Returns

GRNMatrix: pandas.DataFrame: edge information of the GRN will be returned as a DxD matrix

References

grnet.abstract.Estimator: https://grnet.readthedocs.io/en/latest/grnet.abstract.html#grnet.abstract.Estimator

class grnet.models.PC(data: DataFrame, n: int | None = None, random_state: int = 0)

Bases: Estimator

pgmpy wrapper class for PC algorithm

Methods

__init__(: self, data: pandas.core.frame.DataFrame, n: int, random_state: int
) -> None:: initialize attributes
estimate(: self, variant: str, ci_test: str, max_cond_vars: int, return_type: str, significance_level: float, n_jobs: int, show_progress: bool
) -> None:: estimate network and save edges (a list of tuples) as self.edges by PC algorithm actual implementation of PC algorithm is a wrapper of pgmpy.estimators.PC default values are altered from the original codes for some arguments to adjust for GRNs
get_matrix(: self
) -> pandas.core.frame.DataFrame:: export network information as DxD matrix of 0 or 1 elements

Attributes

data: pandas.core.frame.DataFrame: input data or resampled data (data will be resampled if n is specified in self.__init__)
model: pgmpy.estimators.PC.PC: model information (for debugging)
edges: List[tuple]: information of edges are saved as a list of tuples after self.estimate was run

References

pgmpy.estimators.PC: https://pgmpy.org/structure_estimator/pc.html?highlight=pc

estimate(variant: str = 'stable', ci_test: str = 'pearsonr', max_cond_vars: int | None = None, return_type: str = 'dag', significance_level: float = 0.01, n_jobs: int = -1, show_progress: bool = False) → None

Parameters

variant: str, default: “stable”, ci_test: str, default: “pearsonr”, max_cond_vars: int, default: None, return_type: str, default: “dag”, significance_level: float, default: 0.01, n_jobs: int, default: -1, show_progress: bool, default: False

Returns

None

References

pgmpy.estimators.PC: https://pgmpy.org/structure_estimator/pc.html?highlight=pc
grnet.abstract.Estimator: https://grnet.readthedocs.io/en/latest/grnet.abstract.html#grnet.abstract.Estimator

get_matrix() → DataFrame

Parameters

None

Returns

GRNMatrix: pandas.core.frame.DataFrame: edge information of the GRN will be returned as a DxD matrix

References

grnet.abstract.Estimator: https://grnet.readthedocs.io/en/latest/grnet.abstract.html#grnet.abstract.Estimator

class grnet.models.PretrainedModel(data: DataFrame, n: int | None = None, random_state: int = 0)

Bases: Estimator

class for pretrained GRN matrix (to be treated as a subclass of Estimator)

Methods

__init__(: self, data: pandas.DataFrame, n: int, random_state: int
) -> None:: initialize attributes
estimate(: self
) -> None:: always ignored (preserved for consistency)
get_matrix(: self
) -> pandas.DataFrame:: returns the original GRN matrix

Attributes

data: pandas.DataFrame: GRN matrix
edges: List[tuple]: information of edges are saved as a list of tuples

References

grnet.abstract.Estimator: https://grnet.readthedocs.io/en/latest/grnet.abstract.html#grnet.abstract.Estimator

estimate() → None: Parameters

None

Returns

None

get_matrix() → DataFrame

Parameters

None

Returns

GRNMatrix: pandas.DataFrame: edge information of the GRN will be returned as a DxD matrix