|
Abstract
|
One of the most challenging issues in the time series analysis is the detection of fractal and multifractal properties. In spite of introducing many methods for this problem, to find new methods for extracting some important properties such as the strength of the nonlinear correlation is still noteworthy. In this article, we try to introduce a novel method for detecting nonlinear correlations in an arbitrary time series, by using the visibility algorithm and the graph theory. We show that some parameters of mapped networks such as the average degree, the standard deviation and the maximum eigenvalue of the adjacency matrix can be used as quantities for measuring the strength of nonlinear correlations.
|