C38 - Multiple or Simultaneous Equation Models: Classification Methods; Cluster Analysis; Principal Components; Factor ModelsReturn

At present, enterprises are forced to serve their customers as quickly as possible if they want to succeed on turbulent global markets. Enterprises are looking for regions with high-quality infrastructure where they can establish new logistics centres that enable enterprises to serve their customers quickly. This paper focuses on the segmentation of the European Union market for enterprises that are willing to set up logistics centres in order to be able to distribute products fluently and more quickly to their customers in Europe. An agglomerative hierarchical clustering algorithm was used and Ward's criterion applied for the purposes of market segmentation. A Logistic Performance Index and the indicator Dealing with Construction Permits were used as two relevant dimensions reflecting the market attractiveness of identified clusters. Based on the given statistical output, fundamental marketing concepts were formulated for each cluster composed of EU countries with similar characteristics.

Cluster analysis is a method of multivariate data processing. It can be easily applied, for example, by means of statistical software. The article focuses on an application of cluster analysis in the field of taxation. The possibility of applying the method is compared for each tax or groups of taxes at two levels: by an individual state and between selected countries. It is also possible to determine, on the basis of cluster analysis, the impact of taxes on economic operators, which can be sorted according to various criteria (income, types of households represented by their members, etc.). An overview of studies in which the selected method of cluster analysis applied to selected datasets shows that cluster analysis has been supplemented by another statistical method or that the clustering process is repeated on a selected part of a set of objects or characters. The reason is usually better interpretation of results and complementary broader context.

The aim of the paper is to estimate models for household classification from the point of view of their financial situation. The models are constructed on the basis of data from the Living Conditions 2010 survey. The target indicators are the possibility of a household to afford a week-long vacancy outside home, the possibility of a household to afford paying an unplanned expenditure in a certain amount, and an evaluation of how a household is economical with its income. The explanatory indicators are the gender of the head of the household (HOH), the education level of the HOH, the marital status of the HOH, the age of the HOH, and the household type according to the OECD classification. For this purpose, classification trees and logistic regression were applied. The models obtained were evaluated according to the total success rate and the F-measure. The education level of the head of the household was found to be the most important indicator for the prediction.

Questionnaire survey data are usually characterized by a great amount of ordinal variables. For multivariate analysis, it is suitable to reduce task dimensionality. The aim of this paper is a comparison of the results obtained by the analysis of data files with ordinal variables using selected methods for dimensionality reduction. The results are in the form of individual component values (e.g. factor loadings). For better interpretation and comparability, these component values were consequently analyzed by fuzzy clustering. On the basis of the obtained clusters of variables, we determined the optimal number of dimensions. We applied silhouette and Dunn's partition coefficients. Furthermore, we tried to merge the results received by individual methods on the basis of the sCSPA technique (soft version of cluster-based similarity partitioning algorithm). We considered groups of different methods and searched the best solution. The problems are illustrated by means of two real data files obtained from questionnaire surveys. We used SPSS, STATISTICA, Latent GOLD and S-PLUS systems.