Fit finite mixture distribution models to grouped data and conditional data by maximum likelihood using a combination of a Newton-type algorithm and the EM algorithm.
