Fits mixtures of multivariate t-distributions (with eigen-decomposed covariance structure) via the expectation conditional-maximization algorithm under a clustering or classification paradigm.
