Título: Identifying Nonlinearity in Financial Time Series
Autores: Lapenta, Evangelina S.; Abecasis, Sara M.
Resumo: The nonlinearity of financial time series requires the application of different tests to identify it, some of them coming out from outside financial field. Different outlooks of such tests appear scattered in the literature. We give some general essential considerations on nonlinear systems and then we describe three tests: the rescaled range (R/S) analysis for the estimation of the Hurst exponent, the Brock, Dechert and Scheinkman (BDS) statistics, and the estimation of the largest Lyapunov exponent. The aim of the present work is to present a review of the latest versions of these tests and to compare them by applying to the S&P 500 index.
Código DOI: 10.21528/CBRN2001-022
Artigo em pdf: 5cbrn_022.pdf
Arquivo BibTex: 5cbrn_022.bib