mrtool.evidence_score package¶
scorelator¶
-
class
Scorelator(ln_rr_draws, exposures, exposure_domain=None, ref_exposure=None, score_type='area')[source]¶ Bases:
objectEvaluate the score of the result. Warning: This is specifically designed for the relative risk application. Haven’t been tested for others.
-
static
annotate_between_curve(annotation, x, y_lower, y_upper, ax, mark_area=False)[source]¶ Annotate between the curve.
Parameters: - annotation (str) – the annotation between the curve.
- x (np.ndarray) – independent variable.
- y_lower (np.ndarray) – lower bound of the curve.
- y_upper (np.ndarray) – upper bound of the curve.
- ax (Axes) – axis of the plot.
- mark_area (bool, optional) – If True mark the area. Default to False.
-
get_evidence_score(lower_draw_quantile=0.025, upper_draw_quantile=0.975, path_to_diagnostic=None)[source]¶ Get evidence score.
Parameters: - lower_draw_quantile (float, optional) – Lower quantile of the draw for the score.
- upper_draw_quantile (float, optioanl) – Upper quantile of the draw for the score.
- path_to_diagnostic (Union[str, Path, None], optional) – Path of where the picture is saved, if None the plot will not be saved. Default to None.
Returns: Evidence score.
Return type: float
-
static
-
area_between_curves(lower, upper, ind_var=None, normalize_domain=True)[source]¶ Compute area between curves.
Parameters: - lower (np.ndarray) – Lower bound curve.
- upper (np.ndarray) – Upper bound curve.
- ind_var (Union[np.ndarray, None], optional) – Independent variable, if None, it will assume sample points are evenly spaced. Default to None.
- normalize_domain (bool, optional) – If True, when ind_var is None, will normalize domain to 0 and 1. Default to True.
Returns: Area between curves.
Return type: float