Some results concerning off-training-set and IID error for the Gibbs and the Bayes optimal generalizers

In this paper we analyse the average behaviour of the Bayes-optimal and Gibbs learning algorithms. We do this both for off-training-set error and conventional IID (independent identically distributed) error (for which test sets overlap with training sets). For the IID case we provide a major extension to one of the better known results. We also show that expected IID test set error is a non-increasing function of training set size for either algorithm. On the other hand, as we show, the expected off-training-set error for both learning algorithms can increase with training set size, for non-uniform sampling distributions. We characterize the relationship the sampling distribution must have with the prior for such an increase. We show in particular that for uniform sampling distributions and either algorithm, the expected off-training-set error is a non-increasing function of training set size. For uniform sampling distributions, we also characterize the priors for which the expected error of the Bayes-optimal algorithm stays constant. In addition we show that for the Bayes-optimal algorithm, expected off-training-set error can increase with training set size when the target function is fixed, but if and only if the expected error averaged over all targets decreases with training set size. Our results hold for arbitrary noise and arbitrary loss functions.     

Stigma Management Among Gay/Bisexual Men with HIV/AIDS

People with HIV infection are subjected to prejudice, discrimination and hostility related to the stigmatization of AIDS. To manage the stigma of their disease, they mount complex coping strategies. This paper reports results from a qualitative study that examined gay/bisexual men's experiences of living with HIV infection. Unstructured interviews from a diverse sample of 139 men were analyzed to examine how men coped with AIDS-related stigma. We discerned a variety of stigma management strategies that could be arranged along a continuum from reactive to proactive based on the extent to which they implicitly accepted or challenged the social norms and values that underlie the stigmatization of HIV/AIDS. Reactive strategies to cope with stigma involve defensive attempts to avoid or mitigate the impact of stigma, but imply acceptance of the underlying social norms and values that construct the stigma. Examples of reactive strategies include hiding one's HIV status, presenting one's illness as a less stigmatizing one (e.g., cancer), or distancing one's self from more damaging aspects of AIDS-stigma (e.g., attributing infection to blood transfusion). Proactive strategies challenge the validity of the stigma and imply disavowal and resistance of the social norms and values that underlie the stigma. Examples of proactive strategies include engaging in public educational efforts that address misperceptions about HIV transmission and social activism to change the social and political conditions that affect PWA/HIV.     

Absence behaviour, turnover and disability: A study among city bus drivers in the Netherlands

Absenteeism, turnover and disability, and relationships between them have been studied among city bus drivers in the Netherlands. The theoretical framework for the study was a cumulative process model of work and health. Absenteeism among these drivers was two to three times as high as the national average, while the risk of disablement was more than twice as high as the risk for male Dutch civil servants in general. Bus drivers who had to leave their job for medical reasons did so at a younger age than other groups of civil servants. The main conditions leading to disablement related to the back, tendons and joints (35%), mental disorders (35%) and cardiovascular diseases (12%). It was shown that long-term absenteeism is a strong precursor of future disability. The research suggests a work-related process of progressive deterioration of health and well-being over a reasonably long period of time. Driving city buses seems to be an occupation with high risks for health and well-being.     

On the total outer-connected domination in graphs

A set S of vertices of a graph G is a total outer-connected dominating set if every vertex in V(G) is adjacent to some vertex in S and the subgraph induced by V?S is connected. The total outer-connected domination number γ _toc (G) is the minimum size of such a set. We give some properties and bounds for γ _toc in general graphs and in trees. For graphs of order n, diameter 2 and minimum degree at least 3, we show that $\gamma_{toc}(G)\le \frac{2n-2}{3}$ and we determine the extremal graphs.     

Comment “On the role of data,statistics and decisions in a pandemic” by Jahn et al.

AStA Advances in Statistical Analysis - We comment the paper by Jahn et al. (On the role of data, statistics and decisions in a pandemic, 2022).     

Customer satisfaction: Cost driver or value driver? Empirical evidence from the financial services industry

We investigate the relation between customer satisfaction, customer servicing costs, and customer value in a financial services firm. We find that customer satisfaction is positively associated with future customer servicing costs, as well as with customer value. The relation between customer satisfaction and customer value appears non-linear; higher customer satisfaction appears to have a higher return for the most profitable customer segments. Our findings indicate that customer satisfaction is a value driver; however, customer satisfaction is not cost-free and managers have to consider the costs, as well as the benefits, of increasing customer satisfaction.     

Decision Making for Animal Health and Welfare: Integrating Risk‐Benefit Analysis with Prospect Theory

This study integrated risk‐benefit analysis with prospect theory with the overall objective of identifying the type of management behavior represented by farmers’ choices of mastitis control options (MCOs). Two exploratory factor analyses, based on 163 and 175 Swedish farmers, respectively, highlighted attitudes to MCOs related to: (1) grouping cows and applying milking order to prevent spread of existing infection and (2) working in a precautionary way to prevent mastitis occurring. This was interpreted as being based on (1) reactive management behavior on detection of udder‐health problems in individual cows and (2) proactive management behavior to prevent mastitis developing. Farmers’ assessments of these MCOs were found to be based on asymmetrical evaluations of risks and benefits, suggesting that farmers’ management behavior depends on their individual reference point. In particular, attitudes to MCOs related to grouping cows and applying milking order to prevent the spread of mastitis once infected cows were detected were stronger in the risk domain than in the benefit domain, in accordance with loss aversion. In contrast, attitudes to MCOs related to working in a precautionary way to prevent cows from becoming infected in the first place were stronger in the benefit domain than in the risk domain, in accordance with reverse loss aversion. These findings are of practical importance for farmers and agribusiness and in public health protection work to reduce the current extensive use of antibiotics in dairy herds.     

A solution for two-person bargaining problems

Everyday bargaining problems are often solved by tossing a coin. A solution for two-person bargaining problems is axiomatized, which is a Pareto-optimal generalization of this coin tossing method. The super-additive solution of Perles and Maschler is also shown to be a generalization of this method. Various properties of our solution are studied, including continuity and risk sensitivity, and compared with properties of other solutions discussed in the literature.This study was financed by the Yrjö Jahnsson Foundation, which is gratefully acknowledged. I like to thank an anonymous referee for useful comments.     

SIMULTANEOUS DETERMINATION OF CHURCH CONTRIBUTIONS AND CHURCH ATTENDANCE

In a standard model of household allocation of time and income, church attendance and church contributions are determined simultaneously. There are life-cycle aspects to both attendance and contributions, as Azzi and Ehrenberg (1975) and Ehrenberg (1977) have argued, and contributions behavior is affected by tax treatment and by a potential free-rider problem as well. This paper motivates an econometric specification which recognizes these features and estimates the parameters with a data set of over 2000 Protestant church members.