pyhgf.typing.vectorised.LayerParams#

class pyhgf.typing.vectorised.LayerParams(tonic_volatility_vol=None, tonic_volatility=None, tonic_drift=None, autoconnection_strength=None)[source]#

Per-layer static parameters.

Each field is an array with one entry per node in the layer, or None when the field does not apply to the layer’s kind: volatile layers carry tonic_volatility_vol (plus tonic_volatility when the value level’s own tonic volatility is enabled — see DeepNetwork(tonic_volatility=True)), continuous layers carry the other three.

Parameters:

  • tonic_volatility_vol (jax.Array | None) – The tonic (baseline) volatility of the implied internal volatility level (volatile layers only).

  • tonic_volatility (jax.Array | None) – The tonic (baseline) log-volatility \(\omega\) of the node’s own Gaussian random walk. Continuous layers always carry it; volatile layers carry it only when enabled, and None means the value level has no intrinsic volatility at all.

  • tonic_drift (jax.Array | None) – The constant drift \(\rho\) added to the predicted mean at every time step (continuous layers only).

  • autoconnection_strength (jax.Array | None) – The AR(1) coefficient \(\lambda \in [0, 1]\) on the node’s own mean in the prediction; 1.0 is a pure random walk (continuous layers only).

__init__(tonic_volatility_vol=None, tonic_volatility=None, tonic_drift=None, autoconnection_strength=None)#
Parameters:

  • tonic_volatility_vol (Array | None)

  • tonic_volatility (Array | None)

  • tonic_drift (Array | None)

  • autoconnection_strength (Array | None)

Return type:

None

Methods

__init__([tonic_volatility_vol, ...])

create(n_nodes[, tonic_volatility_vol, ...])

Initialise volatile-layer params with defaults.

create_continuous(n_nodes[, ...])

Initialise continuous-layer params with the nodalised defaults.

Attributes

autoconnection_strength

tonic_drift

tonic_volatility

tonic_volatility_vol