numeraire.PricingOutput

class numeraire.PricingOutput(predicted: DataFrame, realized: DataFrame, method: str, config_hash: str, data_vintage: str, run_id: str, protocol: str, capability: str = 'to_pricing', meta: dict[str, ~typing.Any]=<factory>)

Bases: object

Output for a to_pricing method: predicted expected returns vs realized, on test assets.

predicted and realized are both (date x asset): predicted.loc[t] is the model’s cross-section of expected returns for the return realized over (t, t+h] and realized.loc[t] is that realized return (nan for an asset absent / not yet realized at t). protocol records the discipline the panels were produced under — "walk_forward" (per-fold PIT refits, backtest_pricing()) or "in_sample" (one full-sample fit, backtest_pricing_in_sample()) — and flows straight through to every result row so an explanatory in-sample R^2 stays distinguishable from an out-of-sample one.

__init__(predicted: DataFrame, realized: DataFrame, method: str, config_hash: str, data_vintage: str, run_id: str, protocol: str, capability: str = 'to_pricing', meta: dict[str, ~typing.Any]=<factory>) → None

Methods

__init__(predicted, realized, method, ...)

Attributes

capability
universe	Compact universe label (n=<#assets> for panels, the name for a single asset).
predicted
realized
method
config_hash
data_vintage
run_id
protocol
meta

property universe: str

Compact universe label (n=<#assets> for panels, the name for a single asset).