Inference for the generalized Rayleigh distribution based on progressively censored data

In this paper, and based on a progressive type-II censored sample from the generalized Rayleigh (GR) distribution, we consider the problem of estimating the model parameters and predicting the unobserved removed data. Maximum likelihood and Bayesian approaches are used to estimate the scale and shape parameters. The Gibbs and Metropolis samplers are used to predict the life lengths of the removed units in multiple stages of the progressively censored sample. Artificial and real data analyses have been performed for illustrative purposes.     

Sparse conformal predictors

Conformal predictors, introduced by Vovk et al. (Algorithmic Learning in a Random World, Springer, New York, 2005), serve to build prediction intervals by exploiting a notion of conformity of the new data point with previously observed data. We propose a novel method for constructing prediction intervals for the response variable in multivariate linear models. The main emphasis is on sparse linear models, where only few of the covariates have significant influence on the response variable even if the total number of covariates is very large. Our approach is based on combining the principle of conformal prediction with the ℓ ₁ penalized least squares estimator (LASSO). The resulting confidence set depends on a parameter ε>0 and has a coverage probability larger than or equal to 1−ε. The numerical experiments reported in the paper show that the length of the confidence set is small. Furthermore, as a by-product of the proposed approach, we provide a data-driven procedure for choosing the LASSO penalty. The selection power of the method is illustrated on simulated and real data.     

Modeling longitudinal count data with dropouts

This paper explores the utility of different approaches for modeling longitudinal count data with dropouts arising from a clinical study for the treatment of actinic keratosis lesions on the face and balding scalp. A feature of these data is that as the disease for subjects on the active arm improves their data show larger dispersion compared with those on the vehicle, exhibiting an over‐dispersion relative to the Poisson distribution. After fitting the marginal (or population averaged) model using the generalized estimating equation (GEE), we note that inferences from such a model might be biased as dropouts are treatment related. Then, we consider using a weighted GEE (WGEE) where each subject's contribution to the analysis is weighted inversely by the subject's probability of dropout. Based on the model findings, we argue that the WGEE might not address the concerns about the impact of dropouts on the efficacy findings when dropouts are treatment related. As an alternative, we consider likelihood‐based inference where random effects are added to the model to allow for heterogeneity across subjects. Finally, we consider a transition model where, unlike the previous approaches that model the log‐link function of the mean response, we model the subject's actual lesion counts. This model is an extension of the Poisson autoregressive model of order 1, where the autoregressive parameter is taken to be a function of treatment as well as other covariates to induce different dispersions and correlations for the two treatment arms. We conclude with a discussion about model selection. Published in 2009 by John Wiley & Sons, Ltd.     

Maximum likelihood approach to vote aggregation with variable probabilities

The Condorcet-Kemeny-Young statistical approach to vote aggregation is based on the assumption that voters have the same probability of comparing correctly two alternatives and that this probability is the same for any pair of alternatives. We relax the second part of this assumption by letting the probability of comparing correctly two alternatives be increasing with the distance between two alternatives in the allegedly true ranking. This leads to a rule in which the majority in favor of one alternative against another one is given a larger weight the larger the distance between the two alternatives in the true ranking, i.e., the larger the probability that the voters compare them correctly. This rule is not Condorcet consistent and does not satisfy local independence of irrelevant alternatives. Yet, it is anonymous, neutral, and paretian. It also appears that its performance in selecting the alternative most likely to be the best improves with the rate at which the probability increases.We would like to thank Michel Le Breton for his encouragement to examine this question and for his comments, as well as Philippe De Donder, Jean-Yves Duclos, Stephen Gordon, Cyril Téjédo and an anonymous referee for their comments.     

Measuring Sharedness of Team-Related Knowledge: Design and Validation of a Shared Mental Model Instrument

Many researchers have explored how people share and construct similar mental models in teams. They have claimed that successful team performance depends on a shared mental model of team members about task, team, equipment and situation. Most of the literature has illustrated simplified relationships between a team's mental model and their performance without a valid instrument addressing the confined and relevant constructs of a shared mental model. This paper describes the instrument development steps and the conceptual framework for factors associated with shared mental models. After development and refinement, the instrument was finalized for use to measure team-related knowledge. The final instrument consists of 42 items that are linked to the five emergent factors of shared mental models including general task and team knowledge, general task and communication skills, attitude toward teammates and task, team dynamics and interactions, and team resources and working environment.     

Impact of having a high-risk pregnancy on future postpartum contraceptive method choice

BackgroundTo compare the knowledge and preference of preconceptional contraception to future postpartum contraceptive method choice in high-risk pregnancies.Research questionDoes a high-risk pregnancy condition affect future postpartum contraceptive method choice?MethodWomen hospitalised at the High Risk Pregnancy unit of a tertiary research and training hospital were asked to complete a self-reported questionnaire that included demographic characteristics, presence of unintended pregnancy, contraceptive method of choice before the current pregnancy, plans for contraceptive use following delivery and requests for any contraceptive counselling in the postpartum period.FindingsA total of 655 pregnant women were recruited. The mean age, gravidity and parity of the women were 27.48 ± 6.25 years, 2.81 ± 2.15 and 1.40 ± 1.77, respectively. High-risk pregnancy indications included 207 (31.6%) maternal, 396 (60.5%) foetal and 52 (7.9%) uterine factors. All postpartum contraceptive choices except for combined oral contraceptives (COCs) usage were significantly different from preconceptional contraceptive preferences (p < 0.001). High-risk pregnancy indications, future child bearing, ideal number of children, income and education levels were the most important factors influencing postpartum contraceptive choices. While the leading contraceptive method in the postpartum period was long-acting reversible contraceptive methods (non-hormonal copper intrauterine device Cu-IUD, the levonorgestrel-releasing intrauterine system (LNG-IUS) (40%), the least preferred method was COCs use (5.2%) and preference of COCs use showed no difference between the preconceptional and postpartum periods (p = 0.202). Overall 73.7% of the women wanted to receive contraceptive counselling before their discharge.ConclusionA high-risk pregnancy condition may change the opinion and preference of contraceptive use, and also seems to affect the awareness of family planning methods.     

Estimation of finite population mean in stratified sampling using scrambled responses in the presence of measurement errors

This study focuses on the estimation of population mean of a sensitive variable in stratified random sampling based on randomized response technique (RRT) when the observations are contaminated by measurement errors (ME). A generalized estimator of population mean is proposed by using additively scrambled responses for the sensitive variable. The expressions for the bias and mean square error (MSE) of the proposed estimator are derived. The performance of the proposed estimator is evaluated both theoretically and empirically. Results are also applied to a real data set.     

A new financial stress index model based on support vector regression and control chart

Financial stress index (FSI) is considered to be an important risk management tool to quantify financial vulnerabilities. This paper proposes a new framework based on a hybrid classifier model that integrates rough set theory (RST), FSI, support vector regression (SVR) and a control chart to identify stressed periods. First, the RST method is applied to select variables. The outputs are used as input data for FSI–SVR computation. Empirical analysis is conducted based on monthly FSI of the Federal Reserve Bank of Saint Louis from January 1992 to June 2011. A comparison study is performed between FSI based on the principal component analysis and FSI–SVR. A control chart based on FSI–SVR and extreme value theory is proposed to identify the extremely stressed periods. Our approach identified different stressed periods including internet bubble, subprime crisis and actual financial stress episodes, along with the calmest periods, agreeing with those given by Federal Reserve System reports.     

The local geographies of welfare in Tunisia: Does neighbourhood matter?*

Amara M, Ayadi M. The local geographies of welfare in Tunisia: Does neighbourhood matter? The aim of this article is to show that spatial analysis techniques outperform non‐spatial statistical counterparts for understanding the geographic determinants of welfare and poverty in Tunisia. First, an Exploratory Spatial Data Analysis, based on a Geographical Information System, was conducted to visualise the local spatial structure of welfare. Second, a spatial autoregressive (SAR) model and a geographically weighted regression (GWR) model, respectively, were used to deal with both spatial autocorrelations and unobserved spatial heterogeneity of households' behaviours. Spatial and non‐spatial models were compared according to their predictive performances. Results of this case study confirm that SAR and GWR spatial models are preferable to the traditional non‐spatial regression model and that they give a better approximation of the Tunisian poverty map.