15 April 2008, Volume 24 Issue 2
    

  • Select all
    |
    学术论文
  • Han Sier, Tian Zheng, Wang Hongjun
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 113-122.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    This paper studies the problem of testing for parameter change in GARCH models. We proposed residual cumulative sum test statistic and obtained the limiting distribution of test statistic under null hypothesis. The results of a simulation study show that the residual cumulative test can offset the drawbacks, such as low powers of the squares cumulative sum test proposed by Kim, Cho and Lee (2000)[1] and the test is also applied to real data analysis.
  • Zhang Dongxia, Liang Hanying
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 123-134.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    Let $\{X_n,n\geq 1\}$ be a strictly stationary sequence of negatively associated random variables with the marginal probability density function $f(x)$. In this paper, we discuss the point asymptotic normality for recursive kernel density
    estimator of $f(x)$.
  • Tu Yujuan, Feng Shiyong
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 135-147.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    According to different cases of normal-distribution process parameters ($\mu,\sigma^2$), several problems are studied in this paper, for monitoring the process with parameters known or unknown, with single-parameter-deviation or with double-parameter-deviation. A series of CUSUM Charts are proposed in standardized or simpler forms mainly based on $Q$-statistics proposed by Quesenberry. In the last part of the paper, the concept of deviation-interested-only (DIO) is put forward and the relevant statistics are given where there is no
    need to know which has deviated, process mean or process variance. The simulation results of Conditional Expected Delay (CED) are shown in each case, and the comparisons in the end of the paper show which new CUSUM charts act well.
  • Yang Chunhua, Peng Zuoxiang
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 148-156.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    Let $\{(X_1(t),\cdots,X_p(t)),0\leq t\leq T\}$ be $p$ dimensional locally stationary Gaussian processes with asymptotically centered mean $m_k(t)$, $k=1,\cdots,p$ and constant variance. $M_k(T)=\sup\{X_k(t),0\leq t\leq T\}$, $k=1,\cdots,p$.
    Under some conditions, the asymptotic distribution of $M(T)=(M_1(T),\cdots,M_p(T))$ as $T\rightarrow\infty$ is obtained.
  • Deng Guohe, Yang Xiangqun
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 157-165.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    This paper is concerned with studying a problem that minimizes the total expected discounted cost over an infinite horizon for a cash management, where the cash fund follows jump-diffusion processes, holding-costs are assumed to be a general quadratic function of the cash level and there exist fixed and proportional transaction costs. Thanks to the variational inequality method in stochastic impulse control theory, we obtain the verification theorem, prove that the optimal control exists, and also gain its mathematical structures.
  • Su Hongsheng, Zhang Youpeng
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 166-174.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    Two-module redundant system's non-Markov model is converted to Markov process model by the use of supplementary variable technique in this paper,
    then according to the proposed model, reliability evaluation of the system life is implemented, and life distribution of the system under no maintenance is also compared, and also the work characteristics under quick maintenance are presented in detail. Application studies in the traction transformer indicate that the analysis approach is quite effective.
  • Dong Ying, Feng Jinghai
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 175-186.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    In this paper we discuss the insurance companies with payment process
    $A_t$ hedge their risk to the level of minimax by buying stocks $S_t$, exchanging foreign -- currency $Q_t$ and buying risk -- free asset $B_t$ in the financial market $(S_t,Q_t,B_t)$. In virtue of Galtchouk-Kunita-Watanabe Decomposition Theorem, the expression of risk is expressed over again. Then we get the hedging strategies of optimization with minimal risk. It gives out a realistic example to apply the important conclusion in this paper, which makes this paper
    to be more practical.
  • Lu Fuzhong
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 187-198.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    On the mixture model $H=\lambda F+(1-\lambda)G$, we derive the consistent estimator $\wh{\lambda}$ of the mixture proportion $\lambda$ under random right censoring, we also discuss the asymptotic normality of the estimator $\wh{\lambda}$.
  • Li Zhiqiang
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 199-207.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    We develop two local quasi-likelihood imputation estimators for mean in a generalized semiparametric models when response variables are missing at random. It is shown that the proposed mean imputation estimators are
    asymptotically normal.
  • Fan Yonghui, Wang Songgui
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 208-216.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    Maximum likelihood estimation is one of the most important estimate methods in the mixed linear models, but to get the maximum likelihood estimates of variance components we must use iterative algorithms in the most cases. The QR decomposition transforms the design matrices into upper-triangle matrix, then we can decrease the orders of the matrices used in the iterative process, and reduce the amount of data in this process. By simulation, we show that QR decomposition can make EM algorithm run much more quickly and get almost the same results we get without QR decomposition. We also study the ANOVA estimation with QR decomposition.
  • Xu Haiyan, Fei Heliang
    CHINESE JOURNAL OF APPLIED PROBABILITY AND STATIST. 2008, 24(2): 217-224.
    Abstract ( ) Download PDF ( ) Knowledge map Save
    In this article, according to the optimal criterion of minimizing the asymptotic variance of the maximum likelihood estimation of the logarithm of the 100$p$th-percentile in the lower tail of the failure-time distribution at work conditions, methods and guidelines are given for planning two-factor time step-stress
    accelerated life test experiments (ALT) for models in which there is no interaction between the factors when life-time of the tested products is exponential distributed at constant testing conditions. The design given here is also proved to be D-optimal.