Learning discrete decomposable graphical models via constraint optimization

Statistical model learning problems are traditionally solved using either heuristic greedy optimization or stochastic simulation, such as Markov chain Monte Carlo or simulated annealing. Recently, there has been an increasing interest in the use of combinatorial search methods, including those based on computational logic. Some of these methods are particularly attractive since they can also be successful in proving the global optimality of solutions, in contrast to stochastic algorithms that only guarantee optimality at the limit. Here we improve and generalize a recently introduced constraint-based method for learning undirected graphical models. The new method combines perfect elimination orderings with various strategies for solution pruning and offers a dramatic improvement both in terms of time and memory complexity. We also show that the method is capable of efficiently handling a more general class of models, called stratified/labeled graphical models, which have an astronomically larger model space.     

Asymptotic Minimax Risk for the White Noise Model on the Sphere

Estimation of an unknown function on the unit sphere of the Euclidean space is considered. The function is observed in Gaussian continuous time white noise. Uniform norm is chosen as a loss function and exact asymptotic minimax risk is derived extending the result of Korostelev (1993). The exact asymptotic minimax risk is also given for the L ₂-loss, applying the result of Pinsker (1980).     

Making the locational choice: A case approach to the development of a theory of offshore outsourcing and internationalization

Offshore outsourcing is one of the most sustained management strategies of the 21st century, and researchers are paying attention to this contemporary, yet complex and controversial phenomenon. Despite the increasing amount of research, from the theoretical and managerial perspectives several important aspects remain inconclusive. The focus in this article is on theory development in two complex areas. The first concerns the location, the idea being to identify the factors that influence the final decision, in which ownership is not the mode of control. Secondly, the aim is to explore whether offshore outsourcing facilitates international expansion. Through a process of theoretical rationalization and inductive case analysis propositions are arrived at according to which non-locational factors most heavily impact the firm's decision to outsource offshore and its subsequent internationalization. The results indicate that this subsequent internationalization may be a by-product of offshore outsourcing, or it may be an intentional strategy. The article thus adds new aspects to the existing theory on outsourcing decision-making, which forms the basis of an emergent theory for future academic research.     

The effect of objective job demands on registered sickness absence spells; do personal,social and job-related resources act as moderators?

Abstract

This study applied Karaseks demand-control model, using sense of coherence (SOC), social support and job control as moderators of effects of job demands on ensuing sickness absence spells in a 3-year follow-up of 856 municipal employees. Among men the results supported the active learning hypothesis. Passive jobs predicted a high number of sickness absence spells and active jobs predicted a low number of spells. for short spells (1-3 days), the demand-control interaction, however, depended on SOC. In active jobs SOC was negatively associated with sickness absence spells; in passive jobs no such association was found. For long spells (3 days), the demand-control interaction depended on occupational level; active jobs were predictive of low absence spells among blue collar men. Among women, the results supported the strain hypothesis. The demand-control interaction, however, depended on household size. In small households, high-strain jobs predicted a high number of spells, while no increase in spells was found in active versus low-strain jobs. In larger households, the number of spells correlated positively with increasing demands even when control was good. The demand-control interaction, however, depended on SOC and spouse support. With strong SOC or spouse support, absence spells in active jobs remained on a relatively low level, otherwise active jobs led to a high number of spells. This suggests that SOC and spouse support may act as protective factors against female role conflicts associated with active jobs     

Subset Selection in Linear Regression using Sequentially Normalized Least Squares: Asymptotic Theory

This article examines the recently proposed sequentially normalized least squares criterion for the linear regression subset selection problem. A simplified formula for computation of the criterion is presented, and an expression for its asymptotic form is derived without the assumption of normally distributed errors. Asymptotic consistency is proved in two senses: (i) in the usual sense, where the sample size tends to infinity, and (ii) in a non‐standard sense, where the sample size is fixed and the noise variance tends to zero.     

Losing More by Losing It: Poker Experience,Sensitivity to Losses and Tilting Severity

In poker, detrimental decision-making as a result of losing control due to negative emotions is known as tilting. Previous evidence suggests that poker experience is related to better emotion regulation in dealing with poker losses, and possibly to reduced severity of tilting in the game. A correlational on-line study (N = 417) was conducted to operationalize the tilting phenomenon by defining certain experiential characteristics that conceivably protect players from tilting or predispose them to it. These characteristics, as well as a measurement of poker experience, were then used in predicting the severity of tilting. It was hypothesized that (1) players with more poker experience are more likely to perceive having tilted less severely, as a result of accumulating poker experience; (2) players with more poker experience have lower severity of tilting; (3) players with more poker experience report lower emotional sensitivity to losses; and (4) players with a higher emotional sensitivity to losses have higher severity of tilting. Hypotheses 1 and 4 were supported, hypothesis 3 was weakly supported, but contrary to hypothesis 2, poker experience was associated with higher tilting severity. It is argued that these results are sensible if experienced players are less likely to tilt in relative terms, per single hand, but more likely to tilt in the long run.     

The effects of outliers on two nonlinearity tests

Two Lagrange multiplier tests for time series nonlinearities in the presence of outliers are examined by simulation experiments. The nonlinearities studied are autoregressive conditional heteroskedasticity (ARCH) and bilinearity; the outlier types are additive, innovative, temporary change and reallocation outliers. The results show that both the sizes and powers of the tests can be severely distorted by even a single outlier. The severity of the distortions depends on the outlier type and magnitude, but also on the underlying process generating 'the series.