Informatica

All throughout the computer science community object-orientation is accepted as being built upon the same basic concepts that human beings use to apprehend reality. This misconception, as we think, is what we focus on in this paper. To show this we analyse two well-recognised object-oriented systems development methods. We try to pinpoint in what way these methods do not correspond to the way human beings apprehend reality in terms of objects. We show that the methods do not use the concepts of object or class in a manner that corresponds to the way human beings use them to apprehend reality. Furthermore the method-creators do not separate the notion of an object and its representation in a model. We also suggest a better adaptation of the searching-for-objects model based on how human beings apprehend reality. When analysing, one should focus on the purpose of the actions and the two different modes with which an object can be treated: present-at-hand and readiness-to-hand. This will increase the conformity between object-orientation and the way human beings apprehend reality.

In this paper we consider parallel numerical integration algorithms for multi-dimensional integrals. A new hyper-rectangle selection strategy is proposed for the implementation of globally adaptive parallel quadrature algorithms. The well known master-slave parallel algorithm prototype is used for the realization of the algorithm. Numerical results on the SP2 computer and on a cluster of workstations are reported. A test problem where the integrand function has a strong corner singularity is investigated. A modified parallel integration algorithm is proposed in which a list of subproblems is distributed among slave processors.

This paper addresses the study of the speech intelligibility enhancement. The speech model, noise sources, perceptual aspects of speech, and performance evaluation are reviewed. The intelligibility enhancement system based on spectral subtraction technique is investigated. Spectral density estimation device based on the algorithm of smoothed periodograms is analysed. Determination of the silence intervals, efficiency of the silence intervals determination, and signal to noise ratio evaluation are discussed. Speech intelligibility enhancement device is described.

In this paper the framework for business object modeling with focus on distributed enterprise is proposed. It is based on Business Object Architecture, UML and Catalysis method. Business Object Architecture is methodology bringing business semantics to component-based development – the next generation of object-oriented methodology. Basic modeling concepts are business objects, business processes and business rules. Process of business process modeling with Business Objects is described and generic modeling patterns are presented. The framework is illustrated via work effort process modeling.

We have studied the design documentation for two industrial software modules to see if they apply ideas corresponding to contracts, as introduced by Bertrand Meyer, either in an intuitive or in a formal way. They did not, and we identified this fact to be a potential risk factor. This paper presents one of the modules studied, consisting of a sequence of switching sections. Starting from this case study, the paper also discusses how switching sections in general can be designed using contracts in order to increase the semantic integrity of the module as a whole.

This paper presents several results associated with the Wright's generalized hypergeometric function, depicting their interesting characterization properties. Special cases are also pointed out.

In this paper two popular time series prediction methods – the Auto Regression Moving Average (ARMA) and the multilayer perceptron (MLP) – are compared while forecasting seven real world economical time series. It is shown that the prediction accuracy of both methods is poor in ill-structured problems. In the well-structured cases, when prediction accuracy is high, the MLP predicts better providing lower mean prediction error.

Structurization of the sample covariance matrix reduces the number of the parameters to be estimated and, in a case the structurization assumptions are correct, improves small sample properties of a statistical linear classifier. Structured estimates of the sample covariance matrix are used to decorellate and scale the data, and to train a single layer perceptron classifier afterwards. In most from ten real world pattern classification problems tested, the structurization methodology applied together with the data transformations and subsequent use of the optimally stopped single layer perceptron resulted in a significant gain in comparison with the best statistical linear classifier – the regularized discriminant analysis.