Parallel and interacting stochastic approximation annealing algorithms for global optimisation

We present the parallel and interacting stochastic approximation annealing (PISAA) algorithm, a stochastic simulation procedure for global optimisation, that extends and improves the stochastic approximation annealing (SAA) by using population Monte Carlo ideas. The efficiency of standard SAA algorithm crucially depends on its self-adjusting mechanism which presents stability issues in high dimensional or rugged optimisation problems. The proposed algorithm involves simulating a population of SAA chains that interact each other in a manner that significantly improves the stability of the self-adjusting mechanism and the search for the global optimum in the sampling space, as well as it inherits SAA desired convergence properties when a square-root cooling schedule is used. It can be implemented in parallel computing environments in order to mitigate the computational overhead. As a result, PISAA can address complex optimisation problems that it would be difficult for SAA to satisfactory address. We demonstrate the good performance of the proposed algorithm on challenging applications including Bayesian network learning and protein folding. Our numerical comparisons suggest that PISAA outperforms the simulated annealing, stochastic approximation annealing, and annealing evolutionary stochastic approximation Monte Carlo.     

Bayesian density regression for discrete outcomes

We develop Bayesian models for density regression with emphasis on discrete outcomes. The problem of density regression is approached by considering methods for multivariate density estimation of mixed scale variables, and obtaining conditional densities from the multivariate ones. The approach to multivariate mixed scale outcome density estimation that we describe represents discrete variables, either responses or covariates, as discretised versions of continuous latent variables. We present and compare several models for obtaining these thresholds in the challenging context of count data analysis where the response may be over‐ and/or under‐dispersed in some of the regions of the covariate space. We utilise a nonparametric mixture of multivariate Gaussians to model the directly observed and the latent continuous variables. The paper presents a Markov chain Monte Carlo algorithm for posterior sampling, sufficient conditions for weak consistency, and illustrations on density, mean and quantile regression utilising simulated and real datasets.     

In Search of Executive Impairment in Pathological Gambling: A Neuropsychological Study on Non-treatment Seeking Gamblers

Pathological gambling is characterized by a persisting maladaptive and recurrent behavior with severe social and psychological consequences. There is evidence of strong comorbidity with psychiatric manifestations as well as cognitive mainly involving executive functions. This study aimed to investigate impairment in executive functions and working memory, and personality traits in a sample of Greek gamblers. Twenty-four men involved in various gambling activities were recruited from ecological settings as probable pathological gamblers. They were assessed with a comprehensive neuropsychological battery involving several executive tasks, the Zuckerman–Kuhlman Personality Questionnaire, the Hospital Anxiety Depression Scale, and the Difficulties in Emotion Regulation Scale. An age- and education-level matched group of 21 men without history of habitual gambling served as controls. As a group, gamblers displayed significantly lower scores on indices of inhibition, decision making and self-reported emotional awareness, and scored higher on impulsivity/sensation seeking personality traits. Notably, gamblers scored similarly or significantly higher on measures of verbal and visuospatial working memory, cognitive flexibility, processing speed, verbal fluency, and sustained attention. Overall, we argue that gamblers do present with specific cognitive deficits, but there is no evidence for a generalized executive impairment, and further stress the importance of investigating cognitive, personality, and psychiatric aspects of gambling on the basis of an ecologically valid sampling.     

On continuity of incomplete preferences

A weak (strict) preference relation is continuous if it has a closed (open) graph; it is hemicontinuous if its upper and lower contour sets are closed (open). If preferences are complete these four conditions are equivalent. Without completeness continuity in each case is stronger than hemicontinuity. This paper provides general characterizations of continuity in terms of hemicontinuity for weak preferences that are modeled as (possibly incomplete) preorders and for strict preferences that are modeled as strict partial orders. Some behavioral implications associated with the two approaches are also discussed.     

Adolescent Online Gambling in Cyprus: Associated School Performance and Psychopathology

This paper presents results from the study of gambling behaviors in a representative sample of Cypriot high-school students. The aim of the study was to ascertain epidemiology of adolescent online gambling in Cyprus and possible correlates. The sample consisted of 2,684 students (48.5 % boys, 51.5 % girls) from the first grades of junior (Gymnasium) and senior (Lyceum) high school. Our results indicate that gamblers presented with lower school achievement and related expectations while scoring consistently higher on measures of Internet addiction, parental bonding and psychopathology. Discriminant analysis revealed two profiles which explained the most variance in gambling behavior; the first profile that corresponded to more explained variance was the student with higher psychopathology, less prosocial behavior, higher Internet addiction score with higher frequencies of online activities and moderate levels of truancy and lower expectation of school achievement. The second profile was that of a student who reported less psychopathology, more prosocial behavior, less involvement with the Internet in general but skipped classes more and his prospects on finishing high school were even slimmer. These results will be utilized in the design of a comprehensive prevention program in an effort to combat online addictive behaviors.     

The Happiness Workout

This study tests for causality from exercise and physical activity to life satisfaction (LS) by applying an instrumental variable approach with the respondents’ perceived benefits of exercise participation as instruments. Using data across 25 countries from the Eurobarometer survey, our results confirm the positive association between exercise and LS. In terms of causality, the results indicate that being active increases LS for both gender, though more for men than women. One main reason for this relationship is that exercise is perceived as being pleasurable, something that policy-makers should keep in mind when designing programmes to get us off the sofa.     

THE OPTIMAL GOVERNMENT SIZE: FURTHER INTERNATIONAL EVIDENCE ON THE PRODUCTIVITY OF GOVERNMENT SERVICES

This paper estimates the optimal government size for several sets of economies by investigating the role of public services in the production process. I assume government services are optimally provided when their marginal product equals unity (the "Barro rule"). The empirical results suggest: (1) government services are significantly productive; (2) they are overprovided in Africa, underprovided in Asia, and optimally provided everywhere else; (3) the optimal government size is 23 percent for the average country but ranges from 14 percent for the average OECD country to 33 percent in South America; and (4) the marginal productivity of government services is negatively related to government size.     

Restricted maximum likelihood estimation of joint mean‐covariance models

The class of joint mean‐covariance models uses the modified Cholesky decomposition of the within subject covariance matrix in order to arrive to an unconstrained, statistically meaningful reparameterisation. The new parameterisation of the covariance matrix has two sets of parameters that separately describe the variances and correlations. Thus, with the mean or regression parameters, these models have three sets of distinct parameters. In order to alleviate the problem of inefficient estimation and downward bias in the variance estimates, inherent in the maximum likelihood estimation procedure, the usual REML estimation procedure adjusts for the degrees of freedom lost due to the estimation of the mean parameters. Because of the parameterisation of the joint mean covariance models, it is possible to adapt the usual REML procedure in order to estimate the variance (correlation) parameters by taking into account the degrees of freedom lost by the estimation of both the mean and correlation (variance) parameters. To this end, here we propose adjustments to the estimation procedures based on the modified and adjusted profile likelihoods. The methods are illustrated by an application to a real data set and simulation studies. The Canadian Journal of Statistics 40: 225–242; 2012 © 2012 Statistical Society of Canada