Grafos de Quantis de Séries Temporais com Incertezas

  • Jonathan G. Ribeiro GCOM - Grupo de Controle e Modelagem,Universidade Federal de São João del-Rei, MG
  • Erivelton G. Nepomuceno Maynooth University, Maynooth
  • Andriana S. L. O. Campanharo Departamento de Bioestatística, Instituto de Biociências, Universidade Estadual Paulista, Botucatu, São Paulo
Keywords: Complex Networks, Time Series, System Identification, Quantile Graphs, Uncertainty

Abstract

The Quantile Graph (QG) method proved to be a promising technique for mapping time series into complex networks, due to the fact that, not only it made possible to use network statistics to characterize time series and time series statistics to characterize networks, but also it is a method with low computational cost, and it is simple to implement. However, in many cases, there is a possibility in which there is some sort of uncertainty in the time series, which implies in some adaptations of the QG method. Based on this matter, the goal of this paper is to analyze the consequences caused by the uncertainties which are present in the time series and to define the impact of these changes in the topology and behavior of the constructed network. Uncertainty is considered by changing the threshold of the quantiles and using Monte Carlo simulation. The results express the network’s metrics values for a time series with different values of uncertainty, mapped using the proposed method. Network’s metrics are compared with no-uncertainty QG scenario. The main results found are that the proposed method is able to express the changes in the constructed network caused by different types of uncertainties.
Published
2022-10-19
Section
Articles