AN ALTERNATE PARTITIONING TECHNIQUE TO QUANTIFY THE REGULARITY OF COMPLEX TIME SERIES

Abstract

In this paper we use the concepts of information theory to analyze the time series obtained from complex systems. The procedure discussed here can be applied to quantify the regularity of chaotic time series, although it might not certify chaos. The main idea is to map the time series into a finite sequence of symbols using an efficient partitioning technique, and quantify the regularity of the resulting sequence by a chosen complexity measure. A proper partitioning technique is essential for any meaningful analysis of the resulting sequence. We have used a clustering technique to partition the time series into a finite sequence and the Lempel–Ziv complexity measure to quantify the regularity of this sequence.