Most, if not all, databases are mixtures of samples from different distributions. In many cases, however, nothing is known about the source components of these mix-tures. Therefore, many methods that