Management

At this time all people, especially managers and businessmen, are exposed to the ever-present information pollution. This is why tools of business intelligence are of great importance; nevertheless the current methods can hardly cope with large and unstructured text sources like World Wide Web that currently becomes more and more important. To achieve this main goal we have to find and verify satisfactorily reliable methods for automatic extraction of a main context of a document, i.e., multidimensional structured characterization representing the main topic of the document. To cope with the multilingual sources we have to develop approaches that would not be dependent on the language of the source and that would not need any additional language dependent tools (like thesauri). In our conception, the context is dynamic – it means that a classification of a document will not be dependent only on the document in question but also on the corpus; the expansion of a corpus can result in a change of a document classification.

This year, the Visegrad countries celebrate together twenty five years since the socio, economic and political changes lead in collapse of totalitarian regimes. But the political transition brought also problems and new challenges not know previously. The first free elections were held in the Visegrad countries that time and their inhabitants could
choose their political representatives. Many of them came originally from the civil society and revolution movements. However, the political decision-making became disconnected from the civil society over time.
We study the present rootedness of local political representatives in the civil society measured by their membership and leadership in civil society organisations in all four Visegrad countries. We are as well interested
in engagement of activists from the civil society in an institutionalised political life, i.e. political parties and political movements.
The purpose of this book is to discuss development of local democracy, local development initiatives and local social capital in growing urban areas in post-transition countries in which the civil society was almost absent or very weak twenty five years ago. This publication comes up with overviews of the situation in four cities, one in each of the Visegrad
countries. First the Polish case of Poznan is discussed. Then three capital cities follow: Czech Prague, Slovakian Bratislava and Hungarian Budapest.

A generalization of our Gaussian distribution mixtures based method to Bidirectional Texture Function (BTF) modelling is discussed. We estimate the local statistical properties of the monospectral version of the fixed view target BTF texture in the form of a Gaussian mixture of product components. The synthesized texture is obtained by means of a step-wise prediction of the whole fixed view BTF texture subspace. In order to achieve an authentic BTF texture and to avoid possible loss of high-frequency spatial details we use optimally chosen pieces of the original BTF measurements in the synthesis phase. In this sense the proposed BTF texture modelling method can be viewed as a statistically controlled sampling.

Annotation : This chapter explains general model-based approaches to several basic pattern recognition applications followed by a concise description of three fundamental multi-dimensional data model classes. For each model class a solution to parameter estimation and model data synthesis is outlined. Finally an overview of the strengths and weaknesses of studied multi-dimensional data model groups is given.

Assuming local and shift-invariant texture properties we describe the statistical dependencies between pixels by a joint probability density of gray-levels within a suitably chosen observation window. We estimate the unknown multivariate density in the form of a Gaussian mixture of product components from data obtained by shifting the observation window. Obviously, the size of the window should be large to capture the low-frequency properties of textures but, on the other hand, the increasing dimension of the estimated mixture may become prohibitive. By considering a subspace approach based on a structural mixture model we can increase the size of the observation window while keeping the computational complexity in reasonable bounds.

A new method of colour texture modelling based on Gaussian distribution mixtures is discussed. We estimate the local statistical properties of the monospectral version of the target texture in the form of a Gaussian mixture of product components. The synthesized texture is obtained by means of a step-wise prediction of the texture image. In order to achieve a realistic colour texture image and to avoid possible loss of high-frequency details we use optimally chosen pieces of the original colour source texture in the synthesis phase. In this sense the proposed texture modelling method can be viewed as a statistically controlled sampling. By using multispectral or mutually registered BTF texture pieces the method can be easily extended also for these textures.

Recently we have proposed Gaussian mixtures as a local statistical model to synthesize artificial textures. We describe the statistical dependence of pixels of a movable window by multivariate Gaussian mixture of product components. The mixture components correspond to different variants of image patches as they appear in the window. In this sense they can be used to identify different segments of the source color texture image. The segmentation can be obtained by means of Bayes formula provided that a proper decomposition of the estimated Gaussian mixture into sub-mixtures is available. In this paper the mixture model is decomposed by maximizing the mean probability of correct classification of pixels into segments in a way taking into account the assumed consistency of final segmentation.

We propose a sequential forward feature selection method to find a subset of features that are most relevant to the classification task. Our approach uses novel estimation of the conditional mutual information between candidate feature and classes, given a subset of already selected features which is utilized as a classifier independent criterion for evaluation of feature subsets. The proposed mMIFS-U algorithm is applied to text classification problem and compared with MIFS method and MIFS-U method proposed by Battiti and Kwak and Choi, respectively. Our feature selection algorithm outperforms MIFS method and MIFS-U in experiments on high dimensional Reuters textual data.

BLAŽEK, L., PUDIL, P., ŠPALEK, J.: The factors aff ecting competitiveness of companies: contribution and limits of the statistical pattern recognition methods. Acta univ. agric. et silvic. Mendel. Brun., 2011, LIX, No. 7, pp. 69-80 The paper elaborates the methodical side of empirical research of factors infl uencing the economic success of companies. The analysis is based on the selective sample of more than 400 stock listed (share holding) companies and limited partnerships located in the Czech Republic. The main goal of the research is to verify, methodically and theoretically, the hypothesis that there is signifi cant mutual dependency between certain types of economic success of companies and a certain typical confi guration of values of selected characteristics which describe these companies. The paper concentrates on an analysis of applying the statistical pattern recognition methodology in the course of verifying this hypothesis. Our analysis confi rms the potential gains connected with the method. Within the sample we identifi ed group of potential factors of competitiveness which can characterize the interdependence between competitiveness and economic performance. statistical pattern recognition, business economics, economic success, fi nancial performance, empirical research Address prof. Ing. Ladislav Blažek, CSc., Katedra podnikového hospodářství, Ekonomicko-správní fakulta,