JOURNAL OF STATISTICAL SCIENCES
Year:2017 | Volume:11 | Issue:1 |Papers Count:10

Consider a repairable system which starts operating at t=0. Once the system fails, it is immediately replaced by another one of the same type or it is repaired and back to its working functions. In this paper, the system's activity is studied from t>0 for a fixed period of time w. Different replacement policies are considered. In each cases, for a fixed period of time w, the probability model and likelihood function of repair process, say window censored, are obtained. The obtained results depend on the lifetime distribution of the original system, so, expression for the maximum likelihood estimator and Fisher information are derived, by assuming the lifetime follows an exponential distribution.

The collective risk model of insurance company with constant initial capital when process of claims number have the poisson distribution with constant rate is considered. For computing the infinite time ruin probability the stochastic processes and differential equations are used. Also a formula is obtained to compute the Lundberg approximation in finding the approximate of infinite time ruin probability based on the distribution function of claims number. The numerical examples to illustrate these results are given and showed that for any value of initial capital the approximate of our infinite time ruin probability is closer to its real value rather than the ruin probability computed by other authors and has less error.

Most of the research of design optimality is conducted on linear and generalized linear models. In applicable studies, in agriculture, social sciences, etc, usually in addition to fixed effects, there is also at least one random effect in the model. These models are known as mixed models. In this article, Beta regression model with a random intercept is considered as a mixed model and locally D-optimal design is calculated for simple and quadratic forms of the model and the trend of changes of optimal design points for different parameter values will be studied. For the simple model, a two point locally D-optimal design has been obtained for different parameter values and in the quadratic model, a three point locally D-optimal design has been acquired. Also, according to the efficiency criterion, these locally D-optimal designs are compared with the same designs. It was observed that the efficiency of optimal design, when the random intercept is not considered in the model is lower than the case in which the random effect is considered.

In this paper the issue of variable selection with new approach in finite mixture of semi-parametric regression models is studying, although it is supposed that data have Poisson distribution. When we use Poisson distribution, two problems such as overdispersion and excess zeros will happen that can affect on variable selection and parameter estimation. Actually parameter estimation in parametric component of the semi-parametric regression model is done by penalized likelihood approach. However, in nonparametric component after local approximation using Teylor series, the estimation of nonparametric coefficients along with estimated parametric coefficients will be calculated. Using new approach leads to a properly variable selection results. In addition to representing related theories, overdispersion and excess zeros are considered in data simulation section and using EM algorithm in parameter estimation leads to increase the accuracy of end results.

In this paper, we introduce a family of bivariate generalized Gompertz-power series distributions. This new class of bivariate distributions contains several models such as: bivariate generalized Gompertz -geometric, -Poisson, - binomial, -logarithmic, -negative binomial and bivariate generalized exponental-power series distributions as special cases. We express the method of construction and derive different properties of the proposed class of distributions. The method of maximum likelihood and EM algorithm are used for estimating the model parameters. Finally, we illustrate the usefulness of the new distributions by means of application to real data sets.

Maximum of the Renyi entropy and the Tsallis entropy are generalization of the maximum entropy for a larger class of Shannon entropy. In this paper we introduce the maximum Renyi entropy and some of the attributes of distributions which have maximum Renyi entropy investigated. The form of distributions with maximum Renyi entropy is power so we state some properties of these distributions and we have a new form of the Renyi entropy. After pointing the topics of minimum Renyi divergence, some other points in this relation have been discussed. An another form of Renyi divergence have also obtained. Therefore we discussed some of the economic applications of the maximum entropy. Meanwhile, the review of the Csiszar information measure, the general form of distributions with minimum Renyi divergence have obtained.

Usually, we estimate the unknown parameter by observing a random sample and using the usual methods of estimation such as maximum likelihood method. In some situations, we have information about the real parameter in the form of a guess. In these cases, one may shrink the maximum likelihood or other estimators towards a guess value and construct a shrinkage estimator. In this paper, we study the behavior of a Bayes shrinkage estimator for the scale parameter of exponential distribution based on censored samples under an asymmetric and scale invariant loss function. To do this, we propose a Bayes shrinkage estimator and compute the relative efficiency between this estimator and the best linear estimator within a subclass with respect to sample size, hyperparameters of the prior distribution and the vicinity of the guess and real parameter. Also, the obtained results are extended to Weibull and Rayleigh lifetime distributions.

We consider the purely sequential procedure for estimating the scale parameter of an exponential distribution, when the risk function is bounded by the known preassigned number. In this paper, we provide explicit formulas for the expectation of the total sample size. Also, we propose how to adjust the stopping variable so that the risk is uniformly bounded by a known preassigned number. In the end, the performances of the proposed methodology are investigated with the help of simulations.

Penalized estimators for estimating regression parameters have been considered by many authors for many decades. Penalized regression with rectangular norm is one of the mainly used since it does variable selection and estimating parameters, simultaneously. In this paper, we propose some new estimators by employing uncertain prior information on parameters. Superiority of the proposed shrinkage estimators over the least absoluate and shrinkage operator (LASSO) estimator is demonstrated via a Monte Carlo study. The prediction rate of the proposed estimators compared to the LASSO estimator is also studied in the US State Facts and Figures dataset.

The (n-k+1)-out-of-n systems are important types of coherent systems and have many applications in various areas of engineering. In this paper, the general inactivity time of failed components of (n-k+1)-out-of-n system is studied when the system fails at time t>0. First we consider a parallel system including two exchangeable components and then using Farlie-Gumbel-Morgenstern copula, investigate the behavior of mean inactivity time of failed components of the system. In the next part, (n-k+1)-out-of-n systems with exchangeable components are considered and then, some stochastic ordering properties of the general inactivity time of the systems are presented based on one sample or two samples.