Search

Text deals with data mining used in libraries. It focuses on „physical“ libraries as they can use data mining in many processes but chooses its scope opposite to digital libraries where data mining has different variety in usage. This is why papers concerning digital libraries are excluded from this summarizing study. All the papers were selected based on concern with Library and Information Science and were divided into several groups. Most of the studies come from academic environment. Papers from czech and slovak environment are described separately. Time limitation for inclusion of the papers is up to year 2015. [ABSTRACT FROM AUTHOR]

The paper focuses on the usage of data-mining techniques as a support tool for decision making. This paper describes the mining of hidden and potentially useful information from databases using data mining methods. These methods, sometimes called as techniques for knowledge discovering, help users, mostly managers, to make qualified decisions in the organization. The aim of the process of knowledge management is not only to collect information, but to transform it into knowledge and use it in a decision-making process. The purpose of this paper was to find and evaluate the different methodological approaches appropriate for customer segmentation. Various data mining techniques were used for demonstration of customer segmentation according to their purchasing behavior within a selected hypermarket. The following techniques were used for clustering: K-means clustering method, Two Step clustering method and Self Organizing Maps. The quality of final models was evaluated by Silhouette measure. It combines the principles of clusters separation and cohesion. Data mining model was constructed from approximately 60 thousand transaction records. Only the food records were selected for the analysis. The paper also examined the effect of the number of dimensions to the clustering. The original variables were reduced into a smaller number of uncorrelated principal components. These components were used for construction of a scatter plot to check the homogeneity of clusters. The results of this analysis confirmed that the reduction of dimensionality is an useful device for the evaluation of generated clusters. [ABSTRACT FROM AUTHOR]