Module stopping 

Stopping rules used for deciding convergence for EM algorithms.

Structs

Both

A stopping rule that lets the EM algorithm run until both the contained stopping rules indicate convergence.

Either

A stopping rule that lets the EM algorithm run until either of the contained stopping rules indicate convergence.

LogLikelihoodTolerance

A stopping rule that lets the EM algorithm run until the absolute difference in successive, normalised log-likelihood values falls below some tolerance.

Steps

A stopping rule that lets the EM algorithm run for a specific number of EM-steps.

WindowLogLikelihoodTolerance

A stopping rule for window EM that lets the algorithm run until the successive sum of normalised block log-likelihood values falls below a certain tolerance.

Traits

Stop

A type capable of deciding whether an EM algorithm should stop.

StoppingRule

A type that can be combined in various ways to create a stopping rule for EM types.