skfolio.descriptor.EarningsChangeToPrice#

class skfolio.descriptor.EarningsChangeToPrice(lag=252)[source]#

Lagged earnings change divided by current market capitalization.

Computes the change in trailing twelve-month net income over a fixed lag, divided by current market capitalization:

\[\text{earnings\_change\_to\_price}(t) = \frac{\text{net\_income\_ttm}(t) - \text{net\_income\_ttm}(t - \text{lag})} {\text{market\_cap}(t)}\]

The first lag observations are NaN because no lagged history is available.

NaNs are allowed as missing observations and propagate when the current, lagged or market-cap value is missing. Non-missing net_income_ttm values must be finite. Non-missing market_cap values must be finite and strictly positive.

This descriptor captures earnings momentum: whether a firm’s profitability is improving or deteriorating relative to its market value [1]. A positive value indicates earnings improvement and a negative value indicates deterioration.

Unlike GrowthRate, which computes x(t) / x(t-lag) - 1, this formulation is well-defined when earnings are negative. A standard growth rate with a negative base produces sign-inverted rankings, making it unsuitable for earnings. By normalizing the level change with market capitalization, the sign of the output reflects the direction of change.

This is a convenience subclass of ChangeToScale with field="net_income_ttm" and scale_field="market_cap".

This descriptor uses aggregate quantities (net income and market capitalization). The per-share equivalent is:

\[\frac{\text{eps\_ttm}(t) - \text{eps\_ttm}(t - \text{lag})} {\text{adj\_close}(t)}\]

The aggregate form is preferred for consistency with the other value descriptors and to avoid split-adjustment mismatches.

Parameters:

lagint, default=252: Number of observations to look back. The interpretation depends on the data frequency: lag=12 means 1 year for monthly data, lag=252 for daily data, lag=4 for quarterly data.

Attributes:

n_assets_int: Number of assets seen during fitting.
asset_names_ndarray of shape (n_assets,): Asset names seen during fitting.
change_to_scale_ndarray of shape (n_assets,): Last earnings change to price value for each asset.

Methods

`fit_transform`(X[, y])	Compute changes in level normalized by current scale.
`get_metadata_routing`()	Get metadata routing of this object.
`get_params`([deep])	Get parameters for this estimator.
`partial_fit_transform`(X[, y])	Compute changes in level normalized by current scale.
`set_params`(**params)	Set the parameters of this estimator.

See also

ChangeToScale: Generic change-to-scale descriptor.
GrowthRate: Simple growth rate for positive-definite characteristics.
EarningsToPrice: Level of trailing earnings to price (value signal).

References

[1]

Fama, E. F., & French, K. R. (2006). “Profitability, investment and average returns.” Journal of Financial Economics, 82(3), 491-518.

Examples

>>> from skfolio.datasets import make_synthetic_characteristics
>>> from skfolio.descriptor import EarningsChangeToPrice
>>>
>>> X = make_synthetic_characteristics()
>>>
>>> descriptor = EarningsChangeToPrice(lag=252)
>>> earnings_change_to_price = descriptor.fit_transform(X)

fit_transform(X, y=None, **fit_params)#

Compute changes in level normalized by current scale.

Parameters:

XAssetPanel: Input panel containing the field and scale_field characteristics configured at construction.
yNone: Ignored. Present for compatibility with scikit-learn’s API.
**fit_paramsdict: Additional fit parameters. Ignored.

Returns:

change_to_scalendarray of shape (n_observations, n_assets): Change in field over the lag window, divided by current scale_field for each observation and asset.

get_metadata_routing()#

Get metadata routing of this object.

Please check User Guide on how the routing mechanism works.

Returns:

routingMetadataRequest: A MetadataRequest encapsulating routing information.

get_params(deep=True)#

Get parameters for this estimator.

Parameters:

deepbool, default=True: If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns:

paramsdict: Parameter names mapped to their values.

partial_fit_transform(X, y=None, **fit_params)#

Compute changes in level normalized by current scale.

This method supports online updates by continuing from the current fitted state. Use fit_transform to start from a clean state.

Parameters:

XAssetPanel: Input panel containing the field and scale_field fields configured at construction.
yNone: Ignored. Present for compatibility with scikit-learn’s API.
**fit_paramsdict: Additional fit parameters. Ignored.

Returns:

change_to_scalendarray of shape (n_observations, n_assets): Change in field over the lag window, divided by current scale_field for each observation and asset.

set_params(**params)#

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters:

**paramsdict: Estimator parameters.

Returns:

selfestimator instance: Estimator instance.