Search

Your search keyword '"Yu, Changjun"' showing total 5 results
Start Over Author "Yu, Changjun" Topic distribution (probability theory)
5 results on '"Yu, Changjun"'

1. Precise large deviations of aggregate claim amount in a dependent renewal risk model.

Author

Liu, Xijun, Yu, Changjun, and Gao, Qingwu

Subjects
*DEVIATION (Statistics), *AGGREGATION (Statistics), *DISTRIBUTION (Probability theory), *ASYMPTOTIC efficiencies, *MATHEMATICAL bounds
Abstract

In this paper, we consider a dependent risk model, in which the claim sizes are ofdependence structure, their inter-arrival times are independent, identically distributed (i.i.d.), and the claim size and its corresponding inter-arrival time satisfy a certain dependence structure described via the conditional distribution of the inter-arrival time given the subsequent claim size being large. We obtain the asymptotics of the lower and upper bounds of precise large deviations for the aggregate amount of claims, which holds uniformly for allxin an infinite interval oft. [ABSTRACT FROM AUTHOR]

Published
2017
Full Text
View/download PDF

2. Tail behavior of supremum of a random walk when Cramér's condition fails.

Author

Yu, Changjun and Wang, Yuebao

Subjects
*RANDOM walks, *DISTRIBUTION (Probability theory), *EXPONENTIAL functions, *MATHEMATICAL convolutions, *LAPLACE transformation
Abstract

We investigate tail behavior of the supremum of a random walk in the case that Cramér's condition fails, namely, the intermediate case and the heavy-tailed case. When the integrated distribution of the increment of the random walk belongs to the intersection of exponential distribution class and O-subexponential distribution class, under some other suitable conditions, we obtain some asymptotic estimates for the tail probability of the supremum and prove that the distribution of the supremum also belongs to the same distribution class. The obtained results generalize some corresponding results of N. Veraverbeke. Finally, these results are applied to renewal risk model, and asymptotic estimates for the ruin probability are presented. [ABSTRACT FROM AUTHOR]

Published
2014
Full Text
View/download PDF

3. Some discussions on the local distribution classes.

Author

Chen, Wei, Yu, Changjun, and Wang, Yuebao

Subjects
*DISTRIBUTION (Probability theory), *EXISTENCE theorems, *SET theory, *MATHEMATICAL analysis, *CHARACTERISTIC functions, *PROBABILITY theory
Abstract

Abstract: In this paper, we put forward and deal with three problems related to local long-tailed distributions and local subexponential distributions. We prove that there exists a local long-tailed distribution that is not locally subexponential, and also present some other relations and differences between the local distribution class and the global distributions class. [Copyright &y& Elsevier]

Published
2013
Full Text
View/download PDF

4. Lower limits and upper limits for tails of random sums supported on

Author

Yu, Changjun, Wang, Yuebao, and Cui, Zhaolei

Subjects
*LIMIT theorems, *STOCHASTIC processes, *RANDOM variables, *DISTRIBUTION (Probability theory), *MATHEMATICAL statistics
Abstract

Abstract: Based on , and , we study the lower limits and upper limits of the quotients of tails as , where is a non-negative integer-valued random variable and is a distribution supported on . Some of the new results, which are different from the corresponding results of , give a positive answer to Problem 2 of under certain conditions. In addition, we give some corresponding results for the local versions and density versions. [Copyright &y& Elsevier]

Published
2010
Full Text
View/download PDF

5. The closure of the convolution equivalent distribution class under convolution roots with applications to random sums

Author

Yu, Changjun, Wang, Yuebao, and Yang, Yang

Subjects
*MATHEMATICAL convolutions, *DISTRIBUTION (Probability theory), *STOCHASTIC processes, *RANDOM variables, *PROBABILITY theory
Abstract

Abstract: Let be a proper distribution on or and be a non-negative integer-valued random variable with masses . Denote . The main result of this paper is that under some suitable conditions, belongs to the convolution equivalent distribution class if and only if belongs to the convolution equivalent distribution class. As applications, some known results on random sums have been extended and improved, which give a positive answer under certain conditions to Problem 1 of . Similarly, some corresponding results for the local distributions and densities have been obtained. [Copyright &y& Elsevier]

Published
2010
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results