run

ExpectationStep.run(state)

Executes the E-step by calculating the responsibility matrix H.

This method computes the log-likelihood of each data point under each component, incorporates the component weights, and normalizes to find the posterior probabilities (responsibilities). The resulting matrix H is then stored in the pipeline state.

Parameters:

state (PipelineState) – The current state of the pipeline, which must contain the input data X and the current mixture model curr_mixture.

Returns:

The updated pipeline state with the H attribute computed and set.

Return type:

PipelineState