论文信息 - On the Mixtures of Probability Distributions

On the Mixtures of Probability Distributions

The problem of mixtures of probability distri- butions in a general situation, where the parameter set and the observational space are measurable spaces, is considered here. A general theory is developed for identifiability of a class of mixing distributions. A characterization theorem of identifia- bility is given and a relationship between the identifiability of the mixing distributions relative to multivariate distributions and to that of corresponding marginals is established. Simpler proofs with lesser restrictions of some well known results are given and are supported by examples. Finally, a general method of estimating the mixing distribution in a metric space is considered. A consistent estimator of the mixing distribution, when the class of mixing distribution is compact in the topology of weak convergence, is constructed.

Satish Chandra | S. Chandra

[1] J. Deely,et al. Construction of Sequences Estimating the Mixing Distribution , 1968 .

[2] S. Yakowitz,et al. On the Identifiability of Finite Mixtures , 1968 .

[3] K. Parthasarathy,et al. Probability measures on metric spaces , 1967 .

[4] Johannes Kerstan,et al. Unbegrenzt teilbare Punktprozesse , 1974 .

[5] J. Kiefer,et al. On the deviations of the empiric distribution function of vector chance variables , 1958 .

[6] H. Teicher. Identifiability of Mixtures , 1961 .

[7] Henry Teicher,et al. On the Convolution of Distributions , 1954 .

[8] J. Grandell. Doubly stochastic Poisson processes , 1976 .

[9] P. Billingsley,et al. Convergence of Probability Measures , 1969 .

[10] S. Yakowitz. A Consistent Estimator for the Identification of Finite Mixtures , 1969 .

[11] H. Robbins. The Empirical Bayes Approach to Statistical Decision Problems , 1964 .

[12] H. Teicher. Identifiability of Finite Mixtures , 1963 .

[13] H. Robbins,et al. Application of the Method of Mixtures to Quadratic Forms in Normal Variates , 1949 .