Estimation Of Parameters In Finite Mixtures Of Distributions From Right Censored Data

  • Ivana Mala University of Economics, Prague
Keywords: Censored Data, Finite Mixture, Parameter Estimation, Simulation

Abstract

In the text the problem of estimation of parameters in finitemixtures of probability distributions in case of the presence of right  censored data is treated.  Only models with known number of components and observable component membership (complete  mixture models) are studied. Possible estimation method of unknown parameters (parameters of components and mixing  proportions)isdescribed. Specific theoretical and.numerical  problems associated with this this type of the modelling are discussed with respect to the mixture models with complete data.A simulation studyis presented toillustrate properties of estimates and sensitivity of results on the proportion of censored  data  and the sample size. For the  simulation10,000 samples with 500 and 1,000 observations from themixtures  with  two  components  of  normal  (symmetricdistribution)  and  lognormal  (asymmetric  distribution)distributions are generated. Results are given in the tablesand  selected  histograms  of  estimates  of  parameters  areshown. All computations are made in the package R.

References

http://dx.doi.org/10.1002/0471721182

http://dx.doi.org/10.1093/biomet/88.2.535

http://dx.doi.org/ 10.1080/02664769922188

http://dx.doi.org/10.1016/j.jspi.2009.06.014

http://dx.doi.org/ 10.1111/j.1467 9469.2006.00501.x

Therneau, T. A Package for Survival Analysis in S. R package version 2.37­2. (2012).

A.F., Makov, U.E. Statistical analysis of

finite mixture distributions, Wiley Series in Probability and Statistics,John Wiley. 1985. www.researchjournals.co.uk

Published
2012-12-01
Section
Articles