Multi-Dimensional Data Aggregation in the Analysis of Self-Similar Processes

Abstract

Analyzing self-similar processes in various fields requires fast and efficient processing of large amounts of data. The frequency and time scalability of self-similar processes require analysis over multiple time periods. Thus it is necessary to develop effective methods of data aggregation. The paper considers the hierarchical organization of time series and multidimensional aggregation based on a graph. The effectiveness of the proposed aggregation methods and their applicability to the analysis of self-similar processes in various fields are evaluated.