Search

This paper presents a model for discovering user preferences from item characteristics. Based on the theory of evidence, the model estimates a probability interval for an item represented by a set of features. This interval represents the item preference within a group of users and it is computed according to preferences expressed in the past. Additionally, a method for moving among different domains and fusing information is outlined. The issue of efficient search subsets of interest within the inclusion lattice is investigated and algorithms are presented. [ABSTRACT FROM AUTHOR]

Abstract: A mathematical framework to model the Bayesian analysis of single-stage decision problems with imprecise utilities is proposed. The main advantage of this model with respect to previous models for such a problem, is that product measurability of the utility function is not necessary, since this model involves iterated expectations instead of an integral over a product space. Conditions for the equivalence between the extensive and normal forms of the Bayesian analysis, within the proposed framework, are obtained. The model is illustrated with an example in which the utility function is not product measurable. [Copyright &y& Elsevier]

In this paper we establish sufficient conditions to guarantee the exchange of the fuzzy expected value for a fuzzy random variable depending on certain parameter and its Hukuhara derivative with respect to the parameter. Main results are obtained for two different sets of general conditions, one of them requiring measurability conditions with respect to the product space the fuzzy random variable is defined from, and the second one just requiring conditions over the projections of the fuzzy random variables. [Copyright &y& Elsevier]

Classification is a long standing problem in computing. There are two broad kinds of classifiers, frequency based and geometry based. Frequency based classifiers often ignore the geometry underlying the data. Conversely, geometry based classifiers take into account the frequency only indirectly. This paper presents a classification algorithm which considers explicitly geometric and statistical characteristics of the data and combines them into a class representation. Reported here are initial experiments with this algorithm using two well known data sets, both with and without noise. The results show that the proposed algorithm is less sensitive to the training data set than other classifiers. [ABSTRACT FROM AUTHOR]