run

MaximizationStep.run(state)

Executes the M-step.

This method iterates through the configured optimization blocks, updates the parameters for each specified component using the appropriate strategy, and then recalculates the mixture weights based on the sum of responsibilities.

Parameters:

state (PipelineState) – The current state of the pipeline. Must contain the responsibility matrix H and the mixture model curr_mixture.

Returns:

The updated pipeline state with the modified curr_mixture. If the H matrix is not available in the input state, the state is returned with an error set, and no modifications are made.

Return type:

PipelineState