ewm_mean

ewm_mean(source, com=None, span=None, halflife=None, alpha=None, min_periods=0, adjust=True, ignore_na=False, name=None, metadata=None)

Apply ewm_mean to long-form panel inputs.

Parameters

source : Panel | Graph

Input numeric Panel or single-output Graph. com : float | None, default None

Center-of-mass decay parameter. Supply exactly one decay parameter. span : float | None, default None

Span decay parameter. Supply exactly one decay parameter. halflife : float | None, default None

Half-life decay parameter. Supply exactly one decay parameter. alpha : float | None, default None

Smoothing or regularization parameter, depending on the operation. min_periods : int, default 0

Minimum number of observations required to produce a value. adjust : bool, default True

Whether to divide by the decaying adjustment factor. ignore_na : bool, default False

Whether missing values are ignored when calculating weights. name : str | None, default None

Optional graph-node name. A generated name is used when omitted. metadata : Mapping[str, Any] | None, default None

Optional metadata stored on the graph node.

Returns

Graph

Lazy single-output graph. Call .compute() to materialize a Panel.

Examples

import polars as pl

from bagelquant_core import Domain, Panel
from bagelquant_core.transformer import ewm_mean

domain = Domain(calendar=["2024-01-02", "2024-01-03", "2024-01-04"], universe=["a", "b"])
source = Panel.from_domain(
    pl.DataFrame({
        "time": ["2024-01-02", "2024-01-03", "2024-01-04"] * 2,
        "asset_id": ["a"] * 3 + ["b"] * 3,
        "value": [1.0, 2.0, 4.0, 2.0, 3.0, 8.0],
    }),
    domain,
)

result = ewm_mean(source, span=2).compute().data
print(result)

Notes

Inputs are long-form panels keyed by (time, asset_id).

Rolling calculations run independently for each asset_id ordered by time.