Developing a Wide Easy-to-Generate Class of Bivariate Copulas

As of late, copulas have drawn great attention in stochastic simulation, financial engineering, and risk management. Their power lies under their ability of modeling dependent random variables. Using a known theorem in probability which proves that the fractional part of the sum of a uniform and an arbitrary independent continuous random variable follows a uniform distribution, we construct a wide class of bivariate copulas in which bivariate random vector generation can be performed easily. Some important members of this new class and their properties together with two invariant correlation measures and some insights in their application are presented.     

How a System Backfires: Dynamics of Redundancy Problems in Security

Increasing attention is being paid to reliability, safety, and security issues in social systems. Scott Sagan examined why more security forces (a redundancy solution) may lead to less security.⁽ 1 ⁾ He discussed how such a solution can backfire due to three major issues (i.e., “common‐mode error,”“social shirking,” and “overcompensation”). In this article, using Sagan's hypotheses, we simulate and analyze a simple and generic security system as more guards are added to the system. Simulation results support two of Sagan's hypotheses. More specifically, the results show that “common‐mode error” causes the system to backfire, and “social shirking” leads to an inefficient system while exacerbating the common‐mode error's effect. Simulation results show that “overcompensation” has no effect of backfiring, but it leads the system to a critical state in which it can easily be affected by the common‐mode error. Furthermore, the simulation results make us question the importance of the initial power of adversaries (e.g., terrorists) as the results show that, for any exogenous level of adversary power, the system endogenously overcompensates to a level that makes the system more susceptible to being attacked.     

Approximate Bayesian analysis of doubly censored samples from mixture of two Weibull distributions

The purpose of the paper is to estimate the parameters of the two-component mixture of Weibull distribution under doubly censored samples using Bayesian approach. The choice of Weibull distribution is made due to its (i) capability to model failure time data from engineering, medical and biological sciences (ii) added advantages over the well-known lifetime distributions such as exponential, Raleigh, lognormal and gamma distribution in terms of flexibility, increasing and decreasing hazard rate and closed-form distribution function and hazard rate. The proposed two-component mixture of Weibull distribution is even more flexible than its conventional form. However, the estimation of the parameters from the proposed mixture is more complex. Further, we have assumed couple of loss functions under non informative prior for the Bayesian analysis of the parameters from the mixture model. As the resultant Bayes estimators and associated posterior risks cannot be derived in the closed form, we have used the importance sampling and Lindley’s approximation to obtain the approximate estimates for the parameters of the mixture model. The comparison between the performances of approximation techniques has been made on the basis of simulation study and real-life data analysis. The importance sampling is found to be better than Lindley’s approximation as it gives better estimation for shape and mixing parameters of the mixture model and computations under this technique are much easier/shorter than those under Lindley’s approximation.     

Deathbed visions from India: a study of family observations in northern Kerala

This is an interview-based study of 104 families and their observations of the last weeks and days of a dying family member. Forty families reported "unusual experiences and behaviors" from the dying person in their last period of life. Thirty of these dying persons displayed behavior consistent with deathbed visions-interacting or speaking with deceased relatives, mostly their dead parents. There were six cases of reported premonitions of death and five possible confusional states with one patient reported to have had both a deathbed vision and confusional experiences. Socio-demographic factors such as gender, age, occupation, or cause and place of death were not found to be significant. Hindu patients appeared to be more likely than Muslim patients to report these experiences. Use of opiates (or not) did not appear to influence reports. The findings are discussed with reference to past studies of deathbed visions as well as their implications for the future pastoral care of dying people and their families.     

Predictors of Life Satisfaction among Urban Iranian Women: An Exploratory Analysis

This paper explores life satisfaction among married and unmarried Iranian women in urban areas. Data is analyzed from a sample of 335 women of which 61% are married and 39% are unmarried (i.e., single, divorced, or widows). Data is collected through the use of self-reported questionnaires administered during 1992 and 1994. A series of path analysis and cross tabulations suggest that for married women life satisfaction is directly linked to their satisfaction with marriage, employment, and their leisure experiences. There is an inverse relationship between satisfaction and the women's activity, however. For unmarried women, satisfaction is affected by their leisure experiences and educational level. The study suggests that any effort to reduce or increase the educational, employment, or leisure activities of women will directly affect women's general satisfaction and therefore affect Iranian society as a whole.     

Are Iranians Happy? A Comparative Study Between Iran and the United States

This study is based on a survey of 544 respondents in 1995 and 1055respondent in 1997 from Tehran, Iran. It measures the degree ofhappiness among Iranians across gender, income, age, and educationallevels. By drawing on satisfaction with various dimensions of life,the study also explores the correlates of happiness and comparesthem with the 1994 GSS survey of Americans.     

On selection of a suitable prior for the posterior analysis of censored mixture of Topp Leone distribution

The paper aims to select a suitable prior for the Bayesian analysis of the two-component mixture of the Topp Leone model under doubly censored samples and left censored samples for the first component and right censored samples for the second component. The posterior analysis has been carried out under the assumption of a class of informative and noninformative priors using a couple of loss functions. The comparison among the different Bayes estimators has been made under a simulation study and a real life example. The model comparison criterion has been used to select a suitable prior for the Bayesian analysis. The hazard rate of the Topp Leone mixture model has been compared for a range of parametric values.