Clustering patient length of stay using mixtures of Gaussian models and phase type distributions
Lalit Kishore Garg, Sally I. McClean, Brian J. Meenan, Elia El‐Darzi, Peter Millard · 2009
Gaussian mixture distributions and Coxian phase type distributions have been popular choices model based clustering of patients' length of stay data. This paper compares these models and presents an idea for a mixture distribution comprising of components of both of the above distributions. Also a mixed distribution survival tree is presented. A stroke dataset available from the English Hospital Episode Statistics database is used as a running example.