English

Cross-correlation of long-range correlated series

Statistical Finance 2009-03-30 v3 Statistical Mechanics Genomics

Abstract

A method for estimating the cross-correlation Cxy(τ)C_{xy}(\tau) of long-range correlated series x(t)x(t) and y(t)y(t), at varying lags τ\tau and scales nn, is proposed. For fractional Brownian motions with Hurst exponents H1H_1 and H2H_2, the asymptotic expression of Cxy(τ)C_{xy}(\tau) depends only on the lag τ\tau (wide-sense stationarity) and scales as a power of nn with exponent H1+H2{H_1+H_2} for τ0\tau\to 0. The method is illustrated on (i) financial series, to show the leverage effect; (ii) genomic sequences, to estimate the correlations between structural parameters along the chromosomes.

Cite

@article{arxiv.0804.2064,
  title  = {Cross-correlation of long-range correlated series},
  author = {Sergio Arianos and Anna Carbone},
  journal= {arXiv preprint arXiv:0804.2064},
  year   = {2009}
}

Comments

14 pages, 8 figures

R2 v1 2026-06-21T10:30:18.885Z