CohortDispositionScenario
Data scenarios · Implementation: mlsysim/engine/scenario_definitions/data.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
A finite cohort of attempts by disposition at a cutoff, its total cost, and the disposition that emits a learning artifact.
Notes
Equation: The calculation members below define attempts, emission_yield, cost_per_attempt, cost_per_artifact.
Assumptions: Dispositions partition a nonempty cohort; artifact yield is positive for finite artifact cost.
Excludes: Overlapping categories, time-varying costs, and causal claims about a yield intervention.
References
Derivation: A finite cohort of attempts by disposition at a cutoff, its total cost, and the disposition that emits a learning artifact. Each member records the corresponding algebra or named physics function.
Examples
Evaluate the named configuration and display a modeled result.
>>> from mlsysim import Scenarios
>>> case = Scenarios.CurationCohortDispositions
>>> value = case.cost_per_artifact()
>>> print(f"{value:.3g~}")
20 USDAttributes
| Name | Description |
|---|---|
| attempts | Total attempt count across all disposition categories. |
| cohort_cost | |
| cost_per_attempt | Cohort cost divided by total attempts, in USD per attempt. |
| dispositions | |
| emission_yield | Dimensionless fraction of attempts in the artifact-emitting disposition. |
| emitted | |
| yield_multiple |
Methods
| Name | Description |
|---|---|
| cost_per_artifact | USD per emitted artifact from calc_cost_per_accepted; an override changes yield only. |
cost_per_artifact
engine.scenario_types.CohortDispositionScenario.cost_per_artifact(
emission_yield=None,
)USD per emitted artifact from calc_cost_per_accepted; an override changes yield only.
Attempt cost over emission yield; another yield prices the same attempts at that yield.
Calculation: mlsysim.physics.agents.calc_cost_per_accepted.
dispositions
Type: Dict[str, int].
Declared default: required.
emitted
Type: str.
Declared default: required.
cohort_cost
Type: Quantity.
Declared default: required.
yield_multiple
Type: Optional[float].
Declared default: None.
attempts
engine.scenario_types.CohortDispositionScenario.attempts
Total attempt count across all disposition categories.
Equation: sum(dispositions.values()).
emission_yield
engine.scenario_types.CohortDispositionScenario.emission_yield
Dimensionless fraction of attempts in the artifact-emitting disposition.
Equation: dispositions[emitted] / attempts.
cost_per_attempt
engine.scenario_types.CohortDispositionScenario.cost_per_attempt
Cohort cost divided by total attempts, in USD per attempt.
Equation: (cohort_cost / attempts).to('USD').