ABSTRACT
At present, it can be listed many different research fields in which modern methods of linear and nonlinear data analysis are successfully used. These methods help at qualitative and/or quantitative assessment of dynamical properties of processes taking place in complex systems. Generally, these may be processes taking place in both models as well as in real-world systems. Indeed, there can be listed a number of examples in different areas ranging from atmosphere and geophysics to processes in physiology and Core of the Internet, etc. Therefore, it is understandable why theoretical and practical aspects of the analysis of complex time series remain as one of the main subjects of interdisciplinary research interests. Nowadays, there are known different conceptual solutions in the complex time series analysis, though problems arise when available real-world measurement datasets do not fulfill strong requirements of contemporary data analysis. In such cases, having at hand not enough long time series of imperfect quality, researchers usually are forced to combine different approaches in order to have at least some understanding on general features of targeted complex processes. Classification analysis, combined with concepts of contemporary complex data analysis, is deemed to be one of such solutions for relatively short datasets.
