Proposal of the methodology for identification of repetitive sequences in big data

Abstract

The aim of this paper is to propose and describe methodology for identification of repetitive sequences in big data sets. These repetitive sequences can represent for example sequences of failures that emerge in industrial processes. Proposed methodology deals with sequences which are based on time, when the elements of particular sequence emerged. One way to approach such identification is to use so called brute-force scanning, but this approach is very demanding on computational power and computational time for big data sets cases. Our methodology approaches this issue from the side of data mining and data analysis point of view.