应用概率统计

 Office Online
 Journal Online

Current Issue

 2009 Vol.25 Issue.4,Published 2009-07-20
 学术论文
 337 A Generalization of the Classical Risk Model Zhao Yongxia,Yin Chuancun In this paper, we extend the classical risk processto the process, which is a spectrally negative L\'evy process minusa subordinator. Then some results of the ruin probability arederived in terms of properties of L\'evy process and some techniquesfrom martingale theory. Finally, we derive some properties ofhitting time and a Pollaczek-Khinchin type formula of the survivalprobability for a spectrally negative L\'evy process. 2009 Vol. 25 (4): 337-344 [Abstract] ( 1813 ) [HTML 0KB] [ PDF 216KB] ( 1435 )
 345 A Kind of Problem of Maximizing the Expected Utility\\from the Terminal Wealth: the Case of Inflation Wang Guangchen,Wu Zhen This paper is concerned with a kind of problemof maximizing the expected utility from the terminal wealth in thecase of inflation. For a class of CRRA utility, we obtain by adirect method an agent's explicit optimal portfolio strategy and thecorresponding maximum expected utility. Moreover, we give someeconomical interpretations. This idea comes from thecompletion-of-square technique of linear quadratic optimal controlproblems. In Section 3, based on the history data of the stock priceand the inflation rate, the values of the parameters in models areestimated with SAS software. And then, the explicit optimalportfolio strategy and the maximum expected utility are also given. 2009 Vol. 25 (4): 345-353 [Abstract] ( 2035 ) [HTML 0KB] [ PDF 245KB] ( 1487 )
 354 Generalized Multivariate Beta Distribution Lu Zhifeng,Wang Juan This paper extends the (generalized)Beta distribution to the (generalized) multivariate Betadistribution. We also study the moment generating function ofmultivariate Beta distribution, obtain the marginal distribution,conditional distribution and regression function of the generalizedmultivariate Beta distribution. 2009 Vol. 25 (4): 354-364 [Abstract] ( 3553 ) [HTML 0KB] [ PDF 237KB] ( 3485 )
 365 EM Algorithm for Haplotype Inference Incorporating\\Ungenotyped Individuals in Pedigree Studies Zhao Hong,Zhu Wensheng,Guo Jianhua Haplotype inference and reconstruction have becomean essential step in linkage analysis and association analysis. Themethods for haplotype inference are all based on the completegenotype data sets. However, genotypes of some individuals in eachpedigree may be partly missing or ungenotyped. In this article, wepropose a new EM algorithm to perform haplotype inferenceincorporating ungenotyped individuals in pedigree studies. We alsogive the standard errors of estimated parameters and evaluate theperformance of our method by simulation studies. 2009 Vol. 25 (4): 365-374 [Abstract] ( 1953 ) [HTML 0KB] [ PDF 268KB] ( 1653 )
 375 Stochastic Prey-Predator Model Zhang Shulin,Wei Zhenghong In this paper, a 2-dimension, nonlinearbirth-death prey-predator model is established. By generatingfunction, we derive the partial differential equations that the meanprocess of our model satisfied and prove that the preys andpredators will extinct with probability 1. 2009 Vol. 25 (4): 375-380 [Abstract] ( 1546 ) [HTML 0KB] [ PDF 176KB] ( 1387 )
 381 On Stochastic Orders for Order Statistics\\from Normal Distributions Huang Yongjun,Zhang Xinsheng In this paper we obtain some new results on stochasticorders for order statistics from normal distributions. Let$X_1,\cdots,X_n,X^*_1,\cdots,X^*_n$ be independent normal randomvariables with $X_{i}\sim N(\mu_i,\sigma^2)$ and $X^*_{i}\simN(\mu^*_i,\sigma^2)$, $i=1,\cdots,n$. Suppose that there exists astrictly monotone function $f$ such that$(f(\mu_{1}),\cdots,f(\mu_{n}))\succeq_{\text{m}}(f(\mu^{*}_{1}),\cdots,f(\mu^{*}_{n}))$,we prove that: (i) if $f'(x)f''(x)\geq 0$, then$X_{(1)}\leq_{\text{st}}X^*_{(1)}$; (ii) if $f'(x)f''(x)\leq 0$,then $X_{(n)}\geq_{\text{st}}X^*_{(n)}$. Moreover, let $X_{i}\simN(\mu,\sigma_i^2)$ and $X^*_{i}\sim N(\mu,\sigma_i^{*2})$,$i=1,\cdots,n$. We obtain that$({1}/{\sigma_{1}},\cdots,{1}/{\sigma_{n}})\succeq_{\text{m}}({1}/{\sigma^{*}_{1}},\cdots,{1}/{\sigma^{*}_{n}})$ implies that$X_{(1)}\leq_{\text{st}}X^*_{(1)}$ and$X_{(n)}\geq_{\text{st}}X^*_{(n)}$. 2009 Vol. 25 (4): 381-388 [Abstract] ( 1956 ) [HTML 0KB] [ PDF 231KB] ( 1316 )
 389 Comparison Theorem for Algebraic Convergence\\of Diffusion Processes Wang Yingzhe The relations of different operators forreversible Markov processes are studied. And some comparisontheorems for the algebraic $L^2$ convergence are obtained. Someoperator can be compared with a special and simple case so that thealgebraic convergence can be got. 2009 Vol. 25 (4): 389-397 [Abstract] ( 1522 ) [HTML 0KB] [ PDF 227KB] ( 1218 )
 398 Maximum Likelihood Estimation of Multivariate t-Model\\with Rao's Simple Structure and Its Exact Distribution Zou Qingming,Wang Jinglong,Zhu Zhongyi In statistical studying, it is always an interestingtopic to obtain the estimation and its distribution for an unknownparameter in model. In this paper, the maximum likelihood estimationof multivariate t-model with Rao's simple structure is considered.By the technique of conditional distribution, the exact distributionof estimators is obtained. 2009 Vol. 25 (4): 398-408 [Abstract] ( 1822 ) [HTML 0KB] [ PDF 243KB] ( 1414 )
 409 On Exact Distribution of a Class of Supremum-type\\Statistics for Goodness of Fit Wei Bocheng:Li Guoying:Zhao Zhiyuan For goodness of fit tests with simple null hypothesis,Zhang (2002) constructed a classes of supremum-type tests. Differentparameter $\lambda$ and different weighted function $q(t)$ result indifferent tests, including the Kolmogorov-Smirov test, Berk andJones (1979) test and so on. So far, only a few tests correspondingto particular $\lambda$ and $q(t)$ have been studied in theliterature. However, for different problems, the best'' tests aredifferent. It is necessary to discuss the tests for all $\lambda$and the general $q(t)$. In this paper, the exact distributions ofthe test statistics for all $\lambda$ and $q(t)\equiv 1$ arederived. When sample size $n$ is large, it takes a long time to getthe exact quantile. So we give some advice on the computationmethods for different sample size by simulation studies, and a realexample to simply illustrate the above methods. 2009 Vol. 25 (4): 409-420 [Abstract] ( 1512 ) [HTML 0KB] [ PDF 252KB] ( 1278 )
 421 Wavelet Estimation in Varying-Coefficient Models Lu Yiqiang,Li Zhilin This paper is concerned with the estimationof varying-coefficient model that is frequently used in statisticalmodeling. The wavelet procedures are developed to estimate thecoefficient functions. The advantage of this approach is to avoidthe restrictive smoothness requirement for nonparametric function ofthe traditional smoothing approaches for varying-coefficient model,such as kernel and local polynomial methods. Furthermore, theconvergence rate of the wavelet estimators is derived and theasymptotic normality is established. Finite sample properties arestudied through Monte Carlo simulations. 2009 Vol. 25 (4): 421-432 [Abstract] ( 1775 ) [HTML 0KB] [ PDF 218KB] ( 1283 )
 433 Bayesian Assessment for Product's Radiation Hardening\\Performance under Lognormal Distribution Han Feng:Wang Jianguo:Qiao DengJiang We summarize three kinds of data oftenoccurred in experiments which are done to estimate the product'sradiation hardening (RH) performance. A method based on Bayesiantheory is proposed in this paper. This method is applied to evaluatethe product's RH performance if the product's RH performance abidesby the lognormal distribution and the samples are Zero-One data. Theprior distributions of unknown parameters are also discussed, andtwo examples are given to illustrate the proposed method. 2009 Vol. 25 (4): 433-440 [Abstract] ( 1701 ) [HTML 0KB] [ PDF 387KB] ( 1334 )
 441 Statistical Analysis on the Differences of Writing Style\\Between First 80 Chapters and Last 40 Chapters in \\《Dream of Red Mansions》\\(An Application of Equivalent Test on Two Independent\\Binomial Populations) Wei Bocheng Based on data analysis, using the theory and method ofequivalent test on two independent binomial populations'', thispaper provides a strong evidence that there are significantdifferences on the writing of sights between first 80 chapters andlast 40 chapters in 《Dream of Red Mansions》. The confidenceprobability of this conclusion is greater than 98\%. 2009 Vol. 25 (4): 441-448 [Abstract] ( 1573 ) [HTML 0KB] [ PDF 251KB] ( 2227 )

 News

 ·
 ·
 ·
 · New and modified homepage for APS comes