Coeficiente de Hurst na comparação de séries de nível do mar

Abstract

The Hurst coefficient is a measure used to determine long memory and similarities in time series, which can be estimated by the R / S Statistics, the Aggregate Variance Method and the Aggregate Variance Differentiation Method. In this work, the Hurst coefficient was applied in monthly series of the level of the but of 10 countries obtained in a period of 26 years, to verify the existence of the long memory effect. In addition, from the estimation of the coefficient to generate clustering to detect the similarity existing between the time series. In view of this, you may note that the following groups were formed: by Norway and New Zealand; Thailand, Malaysia and Hong Kong and the third group Brazil, Alaska, Japan, Australia and Singapore.