pyhgf.typing.vectorised.LayerState#

class pyhgf.typing.vectorised.LayerState(mean, precision, expected_mean, expected_precision, conditional_expected_precision, effective_precision, value_prediction_error, mean_vol, precision_vol, expected_mean_vol, expected_precision_vol, effective_precision_vol, volatility_prediction_error)[source]#

Vectorised per-layer state, as an eqx.Module.

Each field is an array with one entry per node in the layer.

Parameters:

  • mean (jax.Array) – The posterior mean of the value level.

  • precision (jax.Array) – The posterior precision of the value level.

  • expected_mean (jax.Array) – The predicted (expected) mean of the value level.

  • expected_precision (jax.Array) – The marginal predicted precision of the value level.

  • conditional_expected_precision (jax.Array) – The conditional predicted precision of the value level used by the structured-Gaussian (smoothing) update.

  • effective_precision (jax.Array) – The effective precision of the value-level prediction.

  • value_prediction_error (jax.Array) – The value prediction error of the value level.

  • mean_vol (jax.Array) – The posterior mean of the volatility level.

  • precision_vol (jax.Array) – The posterior precision of the volatility level.

  • expected_mean_vol (jax.Array) – The predicted (expected) mean of the volatility level.

  • expected_precision_vol (jax.Array) – The marginal predicted precision of the volatility level.

  • effective_precision_vol (jax.Array) – The effective precision of the volatility-level prediction.

  • volatility_prediction_error (jax.Array) – The volatility prediction error of the volatility level.

__init__(mean, precision, expected_mean, expected_precision, conditional_expected_precision, effective_precision, value_prediction_error, mean_vol, precision_vol, expected_mean_vol, expected_precision_vol, effective_precision_vol, volatility_prediction_error)#
Parameters:

  • mean (Array)

  • precision (Array)

  • expected_mean (Array)

  • expected_precision (Array)

  • conditional_expected_precision (Array)

  • effective_precision (Array)

  • value_prediction_error (Array)

  • mean_vol (Array)

  • precision_vol (Array)

  • expected_mean_vol (Array)

  • expected_precision_vol (Array)

  • effective_precision_vol (Array)

  • volatility_prediction_error (Array)

Return type:

None

Methods

__init__(mean, precision, expected_mean, ...)

create(n_nodes)

Initialise a layer state with defaults.

Attributes

mean

precision

expected_mean

expected_precision

conditional_expected_precision

effective_precision

value_prediction_error

mean_vol

precision_vol

expected_mean_vol

expected_precision_vol

effective_precision_vol

volatility_prediction_error