“…Finite mixture models are popularly and widely used when knowing statistical patterns of primary data within each subpopulation, as well as mixing proportions, produces valuable knowledge that cannot be known from the marginal patterns. For example, finite mixture models extract signals from noisy spectroscopy data by considering signal and noise to be separate subpopulations (Kuss et al, 2002). Also, failure time of a system can be specifically estimated for each possible cause (Mendenhall and Hader, 1958;Papadapoulos and Padgett, 1986).…”