International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems

The paper considers the analytical solution methods of the maximizing entropy or minimizing variance with fixed orness level problems and the maximizing orness with fixed entropy or variance value problems together. It proves that both of these two kinds of problems have common necessary conditions for their optimal solutions. The optimal solutions have the same forms and can be seen as the same OWA (ordered weighted averaging) weighting vectors from different points of view. The problems of minimizing orness problems with fixed entropy or variance constraints and their analytical solutions are proposed. Then these conclusions are extended to the corresponding RIM (regular increasing monotone) quantifier problems, which can be seen as the continuous case of OWA problems with free dimension. The analytical optimal solutions are obtained with variational methods.

This paper investigates the problem of belief update in Bayesian networks (BN) with uncertain evidence. Two types of uncertain evidences are identified: virtual evidence (reflecting the uncertainty one has about a reported observation) and soft evidence (reflecting the uncertainty of an event one observes). Each of the two types of evidence has its own characteristics and obeys a belief update rule that is different from hard evidence, and different from each other. The particular emphasis is on belief update with multiple uncertain evidences. Efficient algorithms for BN reasoning with consistent and inconsistent uncertain evidences are developed, and their convergences analyzed. These algorithms can be seen as combining the techniques of traditional BN reasoning, Pearl's virtual evidence method, Jeffrey's rule, and the iterative proportional fitting procedure.

In this paper, we give an overview of the most representative approaches aimed at querying databases containing ill-known data, starting from the pioneering works by Codd and Lipski and up to very recent proposals. This study focuses on approaches with a clear and sound semantics, based on the notion of possible worlds. Three types of queries are considered: (i) those about attribute values (in an algebraic or SQL-like framework), (ii) those about the properties satisfied by a given set of worlds (i.e., a set of instances of an imprecise database), and (iii) those about the representation of uncertain data. For the first two types, it is emphasized that a trade-off has to be found between expressivity (of the model) and tractability (of the queries in the context of a given model).

In this paper, we propose a novel approach to fast fuzzy modeling based on a new incremental support vector regression (SVR). Firstly a candidate support vectors selection strategy based on kernel Mahalanobis distance measurement is proposed. This strategy is further used to develop a new incremental learning algorithm to speed up the training process of SVR. Then a hybrid kernel function is utilized to represent an SVR model as a TS fuzzy model. Finally a set of fuzzy rules can be directly extracted from the learning results of SVR. Experimental results of two benchmark examples show that the proposed model not only possesses satisfactory accuracy and generalization ability but also costs less computational time.

Two intelligent part-bringing (path planning or finding) algorithms, to bring a part from an initial position to an assembly hole or a receptacle (target or destination) for a purpose of part mating, related to a robotic part assembly task are introduced. These algorithms are then compared through simulations and several criteria. An entropy function, which is a useful measure of the variability and the information in terms of uncertainty, is introduced to measure its overall performance of a task execution related to the part-bringing task. Fuzzy set theory, well-suited to a management of uncertainty, is introduced to address an uncertainty associated with the part-bringing procedure. A degree of uncertainty associated with a part-bringing is used as an optimality criterion, or cost function, e.g. minimum fuzzy entropy, for a specific task execution. It is shown that the machine organizer using a sensor system can intelligently determine an optimal control value, based on explicit performance criteria, to overcome environmental uncertainty. The algorithms use knowledge processing functions such as machine reasoning, planning, inferencing, learning, and decision-making. The results show the effectiveness of the proposed approaches.

Human factors play a key role in many problems, especially those involve human. In this paper, we address the problem of quantifying private information with the consideration of human factors in data publishing for binary sensitive attributes. We first propose five axioms to capture the properties of private information in data publishing. In particular, the fifth axiom considers human factors in the view of social psychology. We believe that a function that quantifies private information has to satisfy all these five axioms. Therefore, we propose an expected gain model, which allows users to tune a weighting factor to reflect the importance of human factors in their applications. The proposed model has been analyzed and compared with some existing models, such as information gain. Moreover, experiments have been performed on real applications to study the practicality of the proposed model.

This section is maintained by Vladik Kreinovich. Please send your abstracts (or copies of papers that you want to see reviewed here) to vladik@utep.edu, or by regular mail to: Vladik Kreinovich, Department of Computer Science, University of Texas at El Paso, El Paso, TX 79968, USA.