This paper studies maximum likelihood estimation of mixing probabilities ρk, k ≥ l, when the data have a density of the form g = Σpkfk for known densities fk. An application to the shock model of Esary, Marshall and Proschan (1973) is considered in some detail. A simulation study which compares the maximum likelihood estimator to other available estimators for the case of finite mixtures is summarized.
Dans cet article nons etudions l'estimation par la methode du maximum de vraisemblance des probabilites pk k ≥ 1, quand l'echantillon a une densitee g = Σpkfk ou les fk sont des densites connues. Une application de cette technique d'estimation au “shoch model” de Esary. Marshall and Proschan 1973 est preesentee. La derniere section contient les grandes lignes d'une etude de Monte Carlo comparant l'estimateur du maximum de vraisemblance avec d'autres estimateurs obtenus si on suppose que seulement un nombre fixe de pk est non nul.
[1]
D. Hosmer.
A Comparison of Iterative Maximum Likelihood Estimates of the Parameters of a Mixture of Two Normal Distributions Under Three Different Types of Sample
,
1973
.
[2]
D. Boes.
On the Estimation of Mixing Distributions
,
1966
.
[3]
P. Billingsley,et al.
Convergence of Probability Measures
,
1969
.
[4]
Shelemyahu Zacks,et al.
The Theory of Statistical Inference.
,
1972
.
[5]
A. W. Marshall,et al.
Shock Models and Wear Processes
,
1973
.
[6]
Bruce M. Hill,et al.
Information for Estimating the Proportions in Mixtures of Exponential and Normal Distributions
,
1963
.
[7]
V. Hasselblad.
Finite mixtures of distributions from the exponential family
,
1969
.
[8]
Keewhan Choi,et al.
An Estimation Procedure for Mixtures of Distributions
,
1968
.