VISUAL ANALYSIS OF TIME-VARYING MULTIDIMENSIONAL DATA SETS

Abstract. This paper contains a description of computational experiments on the application of elastic maps to the analysis of time-varying volumes of textual information. Elastic maps are considered as a tool to provide analytical work with textual information and large information arrays of data. This paper presents the results of numerical experiments on the study of data volumes consisting of frequencies of joint use of words from different parts of speech, for instance “noun + verb” or “adjective + noun”. We consider text collections in Russian for experiments. Previously, static information arrays were mostly considered. It is for them methods of data analysis and methods of visual analytics were developed. Nevertheless, data comes in all the time in various areas of human activity. And in practice it is necessary to know how the cluster picture of multidimensional data volume changes over time. The paper describes the numerical experiments for real time-varying multidimensional data sets. Such experiments allows to analyze the evolution of cluster structure for multidimensional data and to trace the evolution for separate cluster.