Assessing bank soundness with classification techniques

The recent crisis highlighted, once again, the importance of early warning models to assess the soundness of individual banks. In the present study, we use six quantitative techniques originating from various disciplines to classify banks in three groups. The first group includes very strong and strong banks; the second one includes adequate banks, while the third group includes banks with weaknesses or serious problems. We compare models developed with financial variables only, with models that incorporate additional information in relation to the regulatory environment, institutional development, and macroeconomic conditions. The accuracy of classification of the models that include only financial variables is rather poor. We observe a substantial improvement in accuracy when we consider the country-level variables, with five out of the six models achieving out-of-sample classification accuracy above 70% on average. The models developed with multi-criteria decision aid and artificial neural networks achieve the highest accuracies. We also explore the development of stacked models that combine the predictions of the individual models at a higher level. While the stacked models outperform the corresponding individual models in most cases, we found no evidence that the best stacked model can outperform the best individual model.     

DOUBLE LENGTH ARTIFICIAL REGRESSIONS FOR TESTING SPATIAL DEPENDENCE

This paper derives two simple artificial Double Length Regressions (DLR) to test for spatial dependence. The first DLR tests for spatial lag dependence while the second DLR tests for spatial error dependence. Both artificial regressions utilize only least squares residuals of the restricted model and are therefore easy to compute. These tests are illustrated using two simple examples. In addition, Monte Carlo experiments are performed to study the small sample performance of these tests. As expected, these DLR tests have similar performance to their corresponding LM counterparts.     

Logistic discrimination using robust estimators: An influence function approach

Logistic regression is frequently used for classifying observations into two groups. Unfortunately there are often outlying observations in a data set and these might affect the estimated model and the associated classification error rate. In this paper, the authors study the effect of observations in the training sample on the error rate by deriving influence functions. They obtain a general expression for the influence function of the error rate, and they compute it for the maximum likelihood estimator as well as for several robust logistic discrimination procedures. Besides being of interest in their own right, the influence functions are also used to derive asymptotic classification efficiencies of different logistic discrimination rules. The authors also show how influential points can be detected by means of a diagnostic plot based on the values of the influence function     

Minimum variance unbiased invariant estimation of variance components under normality

In a variance components model for normally distributed data, for a specified vector of linear combinations of the variance components, necessary and sufficient conditions are given under which the vector has a uniformly minimum variance unbiased translation-invariant estimator. The competing class of estimators is not restricted to those that are quadratic. For classification models, the conditions are translated into easy-to-check partial balance requirements on the incidence array.     

Block diagrams and splitting criteria for classification trees

Various aspects of the classification tree methodology of Breiman et al., (1984) are discussed. A method of displaying classification trees, called block diagrams, is developed. Block diagrams give a clear presentation of the classification, and are useful both to point out features of the particular data set under consideration and also to highlight deficiencies in the classification method being used. Various splitting criteria are discussed; the usual Gini-Simpson criterion presents difficulties when there is a relatively large number of classes and improved splitting criteria are obtained. One particular improvement is the introduction of adaptive anti-end-cut factors that take advantage of highly asymmetrical splits where appropriate. They use the number and mix of classes in the current node of the tree to identify whether or not it is likely to be advantageous to create a very small offspring node. A number of data sets are used as examples.     

The theory of constraints: a methodology apart?—a comparison with selected OR/MS methodologies

Mingers (J. Oper. Res. Soc. 54 (2003) 559; Int. Trans. Oper. Res. 7 (2000) 673; J. Mingers, A. Gill (Eds.), Multimethodology: Towards the Theory and Practice of Combining Management Science Methodologies, Wiley, Chichester, 1997), Mingers and Brocklesby (Omega—Int. J. Manage. Sci. 25(5) (1997) 489; Systemist 18(3) (1996) 101) and others have sought to develop classificatory frameworks that would be useful in understanding the nature and characteristics of Operational Research/Management Science (OR/MS) methodologies and the philosophical assumptions underpinning them. This paper extends their work to the domain of methods and methodologies known as the Theory of Constraints (TOC). In particular, the paper helps position TOC methods and tools in relation to traditional OR/MS methodologies, methods and tools, and provides a basis for continuing multi-methodological development across the two domains.The paper concludes that the tools, techniques and methods of TOC can be viewed as a methodological set of complementary hard and soft tools and methods that contribute to all phases of activity and across all three social, personal and material dimensions of the Mingers–Brocklesby framework, and share the ontological and epistemological characteristics and assumptions of extant OR/MS methodologies.     

Bayesian feature selection for classification with possibly large number of classes

In what follows, we introduce two Bayesian models for feature selection in high-dimensional data, specifically designed for the purpose of classification. We use two approaches to the problem: one which discards the components which have “almost constant” values (Model 1) and another which retains the components for which variations in-between the groups are larger than those within the groups (Model 2). We assume that p?n, i.e. the number of components p is much larger than the number of samples n, and that only few of those p components are useful for subsequent classification. We show that particular cases of the above two models recover familiar variance or ANOVA-based component selection. When one has only two classes and features are a priori independent, Model 2 reduces to the Feature Annealed Independence Rule (FAIR) introduced by Fan and Fan (2008) and can be viewed as a natural generalization of FAIR to the case of L>2 classes. The performance of the methodology is studies via simulations and using a biological dataset of animal communication signals comprising 43 groups of electric signals recorded from tropical South American electric knife fishes.     

Robust penalized logistic regression with truncated loss functions

The penalized logistic regression (PLR) is a powerful statistical tool for classification. It has been commonly used in many practical problems. Despite its success, since the loss function of the PLR is unbounded, resulting classifiers can be sensitive to outliers. To build more robust classifiers, we propose the robust PLR (RPLR) which uses truncated logistic loss functions, and suggest three schemes to estimate conditional class probabilities. Connections of the RPLR with some other existing work on robust logistic regression have been discussed. Our theoretical results indicate that the RPLR is Fisher consistent and more robust to outliers. Moreover, we develop estimated generalized approximate cross validation (EGACV) for the tuning parameter selection. Through numerical examples, we demonstrate that truncating the loss function indeed yields better performance in terms of classification accuracy and class probability estimation.     

A cross-national and comparative classification of in-country awareness and policy responses to ‘young carers’

The reality for many families where there is chronic illness, mental health problems, disability, alcohol or substance misuse is that children under the age of 18 are involved in caring. Many of these children – known as ‘young carers’ – will be providing regular and significant care, either episodically or over many years, often ‘hidden’ to health, social care and other welfare professionals and services. These children have most often been invisible in social policy and professional practice. What are the reasons why some countries recognize young carers as a priority for social policy while others (most) do not? What are the key factors that influence a country’s awareness and responses to these children? This article provides an original classification and analysis of country-level responses to young carers, drawing on published research, grey literature, policy documents and the authors’ extensive engagement in policy and practice networks for young carers and their families in a wide range of countries. The analysis identifies two of the key factors that influence the extent and nature of these policy responses, focusing on the importance of a reliable in-country research base and the contribution of influential national NGOs and their networks.     

论《太平御览》的文学观念

《太平御览》是一部综合性类书,全书分为55部,编者对各类事物的观念通过引文可以揭示出来。其中,该书的文学观念在各部类所引的诗文中有所体现,而在文部更为集中。该部摘引前代书籍中比较著名的文学理论观点和与“文”有关的史实,在其文体分类与排列上,与前代史书、总集等存在差异,反映出编者独自的文体观念。其批评观尊儒崇圣,详古略今,对诗赋位置的安排有自己的见解,同时还非常重视文体的源流,尽可能追溯各类文体的历史渊源。