Abstract—Classification is one of the most active research and application areas of neural networks. The literature is vast and growing. This paper summarizes the some of the most important developments in neural network classification research. Specifically, the issues of posterior probability estimation, the link between neural and conventional classifiers, learning and generalization tradeoff in classification, the feature variable selection, as well as the effect of misclassification costs are examined. Our purpose is to provide a synthesis of the published research in this area and stimulate further research interests and efforts in the identified topics. Index Terms—Bayesian classifier, classification, ensemble methods, feature variable selection, learning and generalization, misclassification costs, neural networks. I.

Results from a national representative telephone survey of Americans in 2000 show that Internet and mobile phone usage was very similar, and that several digital divides exist with respect to both Internet and mobile phone usage. The study identifies and analyzes three kinds of digital divides for both the Internet and mobile phones—users/nonuser, veteran/recent, and continuing/dropout—and similarities and differences among those digital divides based on demographic variables. The gap between Internet users and nonusers is associated with income and age, but no longer with gender and race, once other variables are controlled. The gap between mobile phone users and nonusers is associated with income, work status, and marital status. The veteran/recent Internet gap is predicted by income, age, education, phone user, membership in community religious organizations, having children, and gender; for mobile phones, age, work status and marital status are predictors. The gap between continuing and dropout users is predicted by education for Internet usage and income for mobile phone usage. Finally, cross-categorization of Internet and mobile phone usage/nonusage is distinguished (significantly though weakly) primarily by income and education. Thus, there are several digital divides, each predicted by somewhat different variables; and while Internet and mobile phone usage levels in 2000 were about the same, their users overlap but do not constitute completely equivalent populations.

ABSTRACT: The therapeutic effects of intercessory prayer (IP) to the Judeo-Christian God, one of the oldest forms of therapy, has had little attention in the medical literature. To evaluate the effects of IP in a coronary care unit (CCU) population, a prospective randomized double-blind protocol was followed. Over ten months, 393 patients admitted to the CCU were randomized, after signing informed consent, to an intercessory prayer group (192 patients) or to a control group (201 patients). While hospitalized, the first group received IP by participating Christians praying outside the hospital; the control group did not. At entry, chi-square and stepwise logistic analysis revealed no statistical difference between the groups. After entry, all patients had follow-up for the remainder of the admission. The IP group subsequently had a significantly lower severity score bas ed on the hospital course after entry (P <.01). Multivariate analysis separated the groups on the basis of the outcome variables (P <.0001). The control patients required ventilatory assistance, antibiotics, and diuretics more frequently than patients in the IP group. These data suggest that intercessory prayer to the Judeo-Christian God has a beneficial therapeutic effect in patients admitted to a CCU.

In this paper, we present a general framework for understanding the role of artificial neural networks (ANNs) in bankruptcy prediction. We give a comprehensive review of neural network applications in this area and illustrate the link between neural networks and traditional Bayesian classification theory. The method of cross-validation is used to examine the between-sample variation of neural networks for bankruptcy prediction. Based on a matched sample of 220 firms, our findings indicate that neural networks are significantly better than logistic regression models in prediction as well as classification rate estimation. In addition, neural networks are robust to sampling variations in overall classi-

Despite the success of many pattern recognition systems in constrained domains, problems that involve noisy input and many classes remain difficult. A promising direction is to use several classifiers simultaneously, such that they can complement each other in correctness. This thesis is concerned with decision combination in a multiple classifier system that is critical to its success. A multiple classifier system consists of a set of classifiers and a decision combination function. It is a preferred solution to a complex recognition problem because it allows simultaneous use of feature descriptors of many types, corresponding measures of similarity, and many classification procedures. It also allows dynamic selection, so that classifiers adapted to inputs of a particular type may be applied only when those inputs are encountered. Decisions by the classifiers are represented as rankings of the class set that are derivable from the results of feature matching. Rank scores contain more ...

There is a trend in recent OCR development to improve system performance by combining recognition results of several complementary algorithms. This thesis examines the classifier combination problem under strict separation of the classifier and combinator design. None other than the fact that every classifier has the same input and output specification is assumed about the training, design or implementation of the classifiers. A general theory of combination should possess the following properties. It must be able to combine anytype of classifiers regardless of the level of information contents in the outputs. In addition, a general combinator must be able to combine any mixture of classifier types and utilize all information available. Since classifier independence is difficult to achieve and to detect, it is essential for a combinator to handle correlated classifiers robustly. Although the performance of a robust (against correlation) combinator can be improved by adding classifiers indiscriminantly, it is generally of interest to achieve comparable performance with the minimum number of classifiers. Therefore, the combinator should have the ability to eliminate redundant classifiers. Furthermore, it is desirable to have a complexity control mechanism for the combinator. In the past, simplifications come from assumptions and constraints imposed by the system designers. In the general theory, there should be a mechanism to reduce solution complexity by exercising non-classifier-specific constraints. Finally, a combinator should capture classifier/image dependencies. Nearly all combination methods have ignored the fact that classifier performances (and outputs) depend on various image characteristics, and this dependency is manifested in classifier output patterns in relation to input imag...

Classification systems play an important role in business decision-making tasks by classifying the available information based on some criteria. The objective of this research is to assess the relative performance of some well-known classification methods. We consider classification techniques that are based on statistical and AI techniques. We use synthetic data to perform a controlled experiment in which the data characteristics are systematically altered to introduce imperfections such as nonlinearity, multicollinearity, unequal covariance, etc. Our experiments suggest that data characteristics considerably impact the classification performance of the methods. The results of the study can aid in the design of classification systems in which several classification methods can be employed to increase the reliability and consistency of the classification.

Abstract not found

This paper analyzes repayment determinants for loans from Caja Los Andes, a Bolivian microlender. The analysis focusses on the influence of recent...

Aquatic macrophytes are non-woody plants, larger than mi-croscopic size, that grow in water. They are an essential component of wetland communities because they provide food and habitat for a variety of wildlife, and they regulate the chemistry of the open water. Unfortunately, they also hinder human activities by clogging reservoirs and affecting recreational activities. Given their impact on environmental processes as well as on human activities, it is important that aquatic macrophytes be monitored and managed wisely. This research focuses on developing a predictive model, based on several biophysical variables, to determine the future distri-bution of aquatic macrophytes. Par Pond, a cooling reservoir at the Savannah River Site in South Carolina, was selected as the study area. Four biophysical variables, including wa-