logml.survival_analysis.artifacts.optimal_cut_off

Classes

OptimalCutOffArtifact(*args, **kwargs)

Object for conducting univariate analysis: optimal cut-off search.

class logml.survival_analysis.artifacts.optimal_cut_off.OptimalCutOffArtifact(*args, **kwargs)

Bases: logml.survival_analysis.artifacts.base.BaseArtifact

Object for conducting univariate analysis: optimal cut-off search.

LABEL = 'optimal_cut_off'

property best_cut_off: float

Returns the best cut-off found.

property best_p_value_fdr: float

Returns the p-value that corresponds to the best cut-off found.

build(container: logml.data.datasets.survival_dataset.UnivariateSurvivalContainer)

Scans through possible cut-off points for a given container, runs logRank test for each and applies FDR to the result p-values.

get_cut_off_points() → List[float]

Returns checked cut-off points.

get_raw_p_values() → List[float]

Returns raw p-values for cut-off points.

get_p_values_fdr() → List[float]

Returns p-values after FDR for cut-off points.

plot() → matplotlib.figure.Figure

Artifact visualization: cut-off points and LogRank tests p-values (after FDR).

column_name: str

threshold: float