`hp_filter` – Hodrick-Prescott filter – trend / cycle decomposition.#

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.hp_filter_transform.

Function signature#

mf.functions.hp_filter_transform(
    panel: pd.DataFrame,
    lambda_: float,
) -> 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.
`lambda_`	`float`	`1600`	> 0	HP smoothing parameter. Convention: 1600 for quarterly, 129600 for monthly (Ravn-Uhlig 2002).

Returns#

pd.DataFrame — scalar result.

Behavior#

statsmodels hpfilter with smoothing parameter params.lamb (1600 for quarterly, 129600 for monthly per Ravn-Uhlig 2002). Returns the cyclical component by default; the trend can also be returned via params.return = 'trend'.

When to use

Extracting business-cycle gaps from trending series.

When NOT to use

Real-time / one-sided forecasting – HP introduces look-ahead bias unless restricted to expanding_window_per_origin.

In recipe context#

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

# Layer L3 recipe fragment
params:
  op: hp_filter

References#

macroforecast design Part 2, L3: ‘feature engineering is a DAG of typed transforms; cascade-depth bounds the longest chain at cascade_max_depth.’
Hodrick & Prescott (1997) ‘Postwar U.S. Business Cycles: An Empirical Investigation’, JMCB 29(1): 1-16.

hp_filter – Hodrick-Prescott filter – trend / cycle decomposition.#