`fourier` – Fourier basis features – sin/cos at fixed harmonics.#

Back to op axis | Back to L3 | Browse all options

Operational op under axis op, sub-layer L3_A_step_op, layer l3. Standalone callable: mf.functions.fourier_transform.

Function signature#

mf.functions.fourier_transform(
    panel: pd.DataFrame,
    n_terms: int,
    period: int,
) -> pd.DataFrame

Parameters#

name	type	default	constraint	description
`panel`	`pd.DataFrame`	—	—	Input panel. Each column is a variable; rows are time periods. Series is promoted to a single-column DataFrame internally.
`n_terms`	`int`	`4`	>= 1	Number of harmonic pairs (sin + cos) to generate. Total output columns: 2 * n_terms.
`period`	`int`	`12`	>= 1	Fundamental period of the seasonal pattern (e.g., 12 for monthly annual cycle, 4 for quarterly).

Returns#

pd.DataFrame — scalar result.

Behavior#

Generates sin/cos pairs at harmonic frequencies of the calendar period (params.period, params.n_harmonics). Captures smooth periodic patterns without the indicator-explosion of season_dummy.

When to use

Smooth seasonality (annual / weekly cycles) where dummies would over-fit.

In recipe context#

Set params.op = "fourier" in the relevant layer to activate this op within a recipe:

# Layer L3 recipe fragment
params:
  op: fourier

References#

macroforecast design Part 2, L3: ‘feature engineering is a DAG of typed transforms; cascade-depth bounds the longest chain at cascade_max_depth.’

fourier – Fourier basis features – sin/cos at fixed harmonics.#