Comparison of Kernel Discriminant Coordinates Algorithms

The construction of kernel discriminant coordinates reduces to the solution of a generalized eigenvalue problem in which both matrices are nonnegative definite. Six different algorithms for solving that problem are described, and the performance of these algorithms is tested on 26 different datasets. The percentage of misclassifications using a linear discriminant function is noted, and the algorithms’ running times are ascertained. Classification is also performed in the space of classical discriminant coordinates.     

The Voice,Language, And The Gendered Self

This paper examines the intersubjective cocreation of gender ambiguities, identifications, and defensive disidentifications as they emerged out of the vicissitudes of an erotic transference–countertransference in the analysis of a man who stutters. The author posits that the patient's ruptured speech provides an entryway into exploring the problematics of gender. For this man, embracing his masculinity placed him in severe conflict with powerful feminine identifications. The author experienced transferentially the embodiment of different gendered states: receptive preoedipal mother, castrating “phallic” mother, sensuous lover, sadistic penetrating male, benevolent father. She suggests that the stuttering can be understood as an emblem of the sacrifice of the lost primordial union with the mother that accompanies the transition into language.     

Subclinical Psychopathic Traits and Romantic Attachment in Treatment-Seeking Couples

Subclinical psychopathic personality traits have been associated with multiple indicators of couple distress and romantic attachment insecurities but essentially in community samples. This study examined dyadic associations between primary and secondary psychopathy and attachment anxiety and avoidance, assessed with self-reported measures, in 183 French-Canadian cohabiting partners seeking couple therapy. Two Actor-Partner Interdependence Models were tested: Model 1 considered psychopathic traits as independent variables and Model 2 treated attachment insecurities as independent variables. Analyses revealed that in both partners, secondary psychopathy was associated with attachment anxiety and avoidance (Models 1 and 2), while primary psychopathy was not only associated with attachment anxiety (Model 2; actor associations). Partner effects between primary and secondary psychopathy in men and women's attachment anxiety were also found in both models. These findings complement those reported in community samples and have clinical implications for treatment-seeking couples.     

How to stop salami science: promotion of healthy trends in publishing behavior

The career of scientists often depends on the number of their published works. This fact leads to the overproduction of low-quality papers burying the important articles and making the knowledge less accessible. One of the methods to counteract these negative aspects might lie in the promotion of healthy trends in publishing behavior. In the present work, the theoretical analysis of a new bibliometric measure is presented, and its potential impact on the publishing strategies is carefully discussed.     

A note on a makespan minimization problem with a multi-ability learning effect

In the scheduling literature the learning effect is perceived as a process of acquiring experience by a processor (e.g. a human worker) in one ability. However, in many real-life problems the processor, during execution of jobs, increases its experience in different, very often independent, abilities (skills). In consequence, it causes the overall growth of the efficiency of the processor. According to this observation, in this paper, we bring into scheduling a new approach called multi-ability learning that generalizes the existing ones and models more precisely real-life settings. On this basis, we focus on a makespan minimization problem with the proposed learning model and provide optimal polynomial time algorithms for its special cases, which often occur in management.     

Estimation in singular partitioned,reduced or transformed linear models

A singular partitioned linear model, i.e. the singular model comprising the main parameters and the nuisance parameters, can be reduced, or transformed to the form in which only linear functions concerning main parameters are involved. In the paper some properties of the best linear unbiased estimators of these functions following from these models are considered.     

Non-minimization of source taxes on labor income: empirical evidence from Germany

The present paper analyzes the choice of tax classes by married couples under the German withholding tax on wage income. Making use of micro data from official tax returns for the year 1998 and conducting simulation analysis, it is shown that more than 20 percent of the observed households do not choose the tax class combination that minimizes withholding. In the second part of the paper, several explanations for this result are discussed. Among those explanations, transaction costs and Pareto-inefficient intrahousehold-decision-making are the most likely.