BinShareShiftScenario

Statistics scenarios · Implementation: mlsysim/engine/scenario_definitions/statistics.py.

Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.

A feature binned at the training quantiles: training shares against recent shares, scored by PSI.

Notes

Equation: PSI = sum((current share - baseline share) * log(current share / baseline share)).

Assumptions: The lists refer to the same bins. Per-bin term access requires positive shares; the aggregate helper defines its own zero handling.

Excludes: Causal drift diagnosis, uncertainty intervals, and a universal PSI alert threshold.

References

Derivation: pSI = sum((current share - baseline share) * log(current share / baseline share)). The individual properties state units and intermediate accounting; this class introduces no fitted coefficient.

Examples

The named scenario supplies illustrative inputs; the output is calculated.

>>> from mlsysim import Scenarios
>>> case = Scenarios.FraudCountryPsi
>>> round(float(case.psi), 3)
0.065

Attributes

Name	Description
baseline_shares
bins
current_shares
psi	Total dimensionless PSI from
terms	Per-bin PSI contributions, each current-minus-baseline share times log of their

Methods

Name	Description
ratio	Current share divided by baseline share for indexed bin `i`.
shares	Pair of baseline and current fractions for the named bin.
term	Dimensionless PSI contribution for the named bin, selected from `terms`.

ratio

engine.scenario_types.BinShareShiftScenario.ratio(i)

Current share divided by baseline share for indexed bin i.

Equation: return current_shares[i] / baseline_shares[i].

engine.scenario_types.BinShareShiftScenario.shares(name)

Pair of baseline and current fractions for the named bin.

Equation: i = bins.index(name); return (baseline_shares[i], current_shares[i]).

term

engine.scenario_types.BinShareShiftScenario.term(name)

Dimensionless PSI contribution for the named bin, selected from terms.

Equation: return terms[bins.index(name)].

baseline_shares

Type: List[float].

Declared default: required.

current_shares

Type: List[float].

Declared default: required.

bins

Type: Optional[List[str]].

Declared default: None.

terms

engine.scenario_types.BinShareShiftScenario.terms

Per-bin PSI contributions, each current-minus-baseline share times log of their ratio.

Equation: return [(c - b) * math.log(c / b) for b, c in zip(baseline_shares, current_shares)].

psi

engine.scenario_types.BinShareShiftScenario.psi

Total dimensionless PSI from physics.statistics.calc_population_stability_index.

Equation: return calc_population_stability_index(baseline_shares, current_shares).