tki.insights.point_insight.OutstandingLastInsight

class tki.insights.point_insight.OutstandingLastInsight(distribution_law: ~typing.Callable[[...], ~numpy.ndarray] = functools.partial(<function power_dist>, beta=0.7), stat_distribution: ~scipy.stats._distn_infrastructure.rv_continuous = <scipy.stats._continuous_distns.norm_gen object>)

By predicting a given distribution for the values sorted descending this insight calculates the likelihood of the last (lowest) value given the later.

The score is calculated by multiplying the impact factor with the p-value.

Parameters

distribution_lawfn(np.ndarray, …) -> np.ndarray

Function describing a distribution to fit the data Defaults to power_dist_fix_beta(0.7)

stat_distributionscipy.stats.rv_continuous

Statistic distribution to describe the distribution of residuals Defaults to scipy.stats.norm

__init__(distribution_law: ~typing.Callable[[...], ~numpy.ndarray] = functools.partial(<function power_dist>, beta=0.7), stat_distribution: ~scipy.stats._distn_infrastructure.rv_continuous = <scipy.stats._continuous_distns.norm_gen object>)

Methods

__init__([distribution_law, stat_distribution])
calc_insight(extraction_result)	Calculate Insight score
plot(result)	Visualizes Insight Result using matplotlib

Attributes

name