Information-Based Evaluation Criterion for Classifier's Performance

In the past few years many systems for learning decision rules from examples were developed. As different systems allow different types of answers when classifying new instances, it is difficult to appropriately evaluate the systems' classification power in comparison with other classification systems or in comparison with human experts. Classification accuracy is usually used as a measure of classification performance. This measure is, however, known to have several defects. A fair evaluation criterion should exclude the influence of the class probabilities which may enable a completely uninformed classifier to trivially achieve high classification accuracy. In this paper a method for evaluating the information score of a classifier's answers is proposed. It excludes the influence of prior probabilities, deals with various types of imperfect or probabilistic answers and can be used also for comparing the performance in different domains.     

Consolidation of Long-Term Memory: Evidence and Alternatives.

Memory loss in retrograde amnesia has long been held to be larger for recent periods than for remote periods, a pattern usually referred to as the Ribot gradient. One explanation for this gradient is consolidation of long-term memories. Several computational models of such a process have shown how consolidation can explain characteristics of amnesia, but they have not elucidated how consolidation must be envisaged. Here findings are reviewed that shed light on how consolidation may be implemented in the brain. Moreover, consolidation is contrasted with alternative theories of the Ribot gradient. Consolidation theory, multiple trace theory, and semantization can all handle some findings well but not others. Conclusive evidence for or against consolidation thus remains to be found. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Extremes of random fields over arbitrary domains with application to concrete rupture stresses

To find the exact probability distribution of the global maximum or minimum of a random field within a bounded domain is a pending problem even for Gaussian fields. Except for very special examples of fields, recourse must be taken to approximate reasoning or asymptotic considerations to be judged with respect to accuracy by simulations. In this paper, the problem is addressed through a functional equation that leads to the definition of a class of distribution functions that depend solely on process or field characteristics and domain quantities that can be calculated explicitly. This distribution function class is studied for Gaussian processes in earlier works by the author and it has been obtained explicitly for Gaussian fields on rectangular domains in the plane. Simulation studies show that rather good predictions are obtained for sufficiently smooth wide band Gaussian processes and fields. In this paper, the distribution function is obtained in general for Gaussian fields over arbitrary bounded domains with piecewise continuous and differentiable boundaries, and as in earlier works the distribution function is tested against empirical distribution functions obtained by simulation of sample functions of a smooth approximately Gaussian field, herein called a broken line Hino field. For completeness this particular field type is defined in appendix a and appendix b. The paper concludes with a statistical application on data for plain concrete tensile strength.     

Dynamique motivationnelle de l'epuisement et du bien-etre chez des enseignants africains.

This study examines the links between antecedents and consequences of Gabonese teachers' (N= 152) work motivation. The theoretical underpinnings are based on the Motivational Model of Job Burnout (MMJB) which is grounded in Self-Determination Theory (SDT) and states that the more supervisors satisfy an employee's need for self-determination, competence, and relatedness, the more self-determined is the latter's work motivation. The level of self-determined motivation has an impact on job satisfaction, which in turn determines life satisfaction. Higher levels of self-determined motivation also buffer emotional exhaustion and psychological distress. The model was confirmed via structural equation modeling and provides support for the validity and generalizability of the MMJB and SDT. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

结构-波浪-海床耦合系统中大圆筒结构的波压力响应

以沉入式大圆筒结构为对象，通过设计随机波浪水槽实验，测取了在单筒和连续筒两种结构型式、两种沉入深度和两种波谱输入下结构表面的动波压力分布。波面高度以及波压力历程均在领域进行分析。通过计算结构表面沿高程和沿环向的波压力谱及波压力传递函数，将单筒实测结果与基于线性绕射理论的解析解进行了比较，同时也详细对比了两种结构型式下的实测结果。文中汇报了在研究中获得的新发现及进行的相应分析，特别是在结构——波浪——海床动力相互作用下对波压力响应机理以及针对连续筒的研究，可为工程应用提供有价值的参考。     

Dynamic stiffness vibration analysis using a high‐order beam model

This work presents the derivation of the exact dynamic stiffness matrix for a high‐order beam element. The terms are found directly from the solutions of the differential equations that describe the deformations of the cross‐section according to the high‐order theory, which include cubic variation of the axial displacements over the cross‐section of the beam. The model has six degrees of freedom at the two ends, one transverse displacement and two rotations, and the end forces are a shear force and two end moments. Using the dynamic stiffness matrix exact vibration frequencies for beams with various combinations of boundary conditions are tabulated and compared with results from the Bernoulli–Euler and Timoshenko beam models. Copyright © 2003 John Wiley & Sons, Ltd.     

Many-Body Terms in van der Waals Cohesion Energy of Nanotubes

We have developed a model for the calculation of van der Waals force for layered systems with axial symmetry. Our result can be applied to compute the cohesion of a carbon nanotube to a substrate, the cohesion between nanotubes, and between shells of multiwall nanotubes. We have obtained unusal power laws for the distance dependence of the many-body van der Waals potential.     

Web-based systems design: a study of contemporary practices and an explanatory framework based on “method-in-action”

This paper reports the findings of a detailed study of Web-based systems design (WBSD) practices in Ireland based on data collected over a 3-year period (2002–2005), the objectives of which were to (1) contribute towards a richer understanding of the current “real-world” context of WBSD by characterising the profile of a typical project (team size, timeframe, nature of requirements, etc.) and identifying the key challenges, constraints, and imperatives (i.e. “mediating factors”) faced by Web-based system designers, and (2) understand how those contextual parameters and mediating factors influence the activity of WBSD as regards the selection and enactment of whatever design practices are therefore engaged (i.e. the use of methods, procedures, etc.). Data was gathered through a survey which yielded 165 usable responses, and later through a series of semi-structured qualitative interviews. Using grounded theory, an explanatory conceptual framework is derived, based on an extension of the “method-in-action” model, the application of which to WBSD has not been previously investigated in depth. It is proposed that this framework of WBSD issues is valuable in a number of ways to educators, researchers, practitioners, and method engineers.     

Multiagent learning using a variable learning rate

Learning to act in a multiagent environment is a difficult problem since the normal definition of an optimal policy no longer applies. The optimal policy at any moment depends on the policies of the other agents. This creates a situation of learning a moving target. Previous learning algorithms have one of two shortcomings depending on their approach. They either converge to a policy that may not be optimal against the specific opponents' policies, or they may not converge at all. In this article we examine this learning problem in the framework of stochastic games. We look at a number of previous learning algorithms showing how they fail at one of the above criteria. We then contribute a new reinforcement learning technique using a variable learning rate to overcome these shortcomings. Specifically, we introduce the WoLF principle, “Win or Learn Fast”, for varying the learning rate. We examine this technique theoretically, proving convergence in self-play on a restricted class of iterated matrix games. We also present empirical results on a variety of more general stochastic games, in situations of self-play and otherwise, demonstrating the wide applicability of this method.     

Appraisal as a Pervasive Determinant of Anger.

L. Berkowitz and E. Harmon-Jones (see record 2004-15096-001) challenge appraisal theories of emotion by describing 2 sets of conditions (physical discomfort and anger-related muscle actions) in which anger appears to be elicited in the absence of theoretically predicted appraisals. In response, the authors discuss the ability of the specific appraisal model they have developed (e.g., C. A. Smith & L. D. Kirby, 2000, 2001; C. A. Smith & R. S. Lazarus, 1990) to account for such instances of anger. First, a number of issues are clarified relevant to the authors' model, including the nature of both the cognitive operations underlying appraisal and the specific appraisals hypothesized to evoke anger. The authors then describe how their model can account for the instances of anger described by L. Berkowitz and E. Harmon-Jones and how both accounts might be tested. (PsycINFO Database Record (c) 2010 APA, all rights reserved)