arch.univariate.recursions_python.VolatilityUpdater.initialize_update¶

abstract VolatilityUpdater.initialize_update(parameters: ndarray, backcast: float | ndarray, nobs: int) → None[source]¶

Initialize the recursion prior to calling update

Parameters:¶

parameters: ndarray¶

The model parameters.

backcast: float | ndarray¶

The backcast value(s).

nobs: int¶

The number of observations in the sample.

Notes

This function is called once per likelihood evaluation and can be used to pre-compute expensive parameter transformations that do not change with each call to update.