|Table of Contents|

A Revised Method to Measure Statistical Dependence Between Two Time Series(PDF)

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

Issue:
2002年03期
Page:
325-329
Research Field:
Publishing date:
2002-06-30

Info

Title:
A Revised Method to Measure Statistical Dependence Between Two Time Series
Author(s):
WangHaiyan LiWen ① ChenWenyan
Department of Applied Mathematics, Southeast University, Nanjing 210096)
Keywords:
t ime series stat istical dependence mutual information correlation integral
PACS:
O212
DOI:
-
Abstract:
A new method to measure stat istical dependence between two random variables is proposed by use of revised generalized mutual informat ion. Based on this method, the formulat ion to calculate statist ical dependence betw een tw o observat ional t ime series w ith ergodicity is obtained by use of correlation integral. By comparing this method w ith the methods that measure stat ist ical dependence using mutual informat ion and generalized mutual information, the advantages of this method are illustrated by two examples.

References:

1   Pompe B. Measuring statistical dependences in a time series. Journal of Statistical Physics, 1993,73( 3/ 4) : 587~ 610
2  Pompe B, Blidh P, Hoyer D, et al. Using mutual information to measure coupling in the cardio respirator y system. IEEE Engineering in Medicine and Biolog y, 1998, 17( 6) : 32~ 39
3  Prichard D, Theiler J. Generalized redundancies for time ser ies analysis. Physica D, 1995, 84:476~ 493
4  Grassberger P, Procaccia I. Estimation of the Kolmorgor ov entropy from a chaotic signal. Phys-ical Review A, 1983, 28: 2 591~ 2 593

Memo

Memo:
-
Last Update: 2002-06-30