Publication: Information-theoretic analysis of seral dependence and cointegration
Loading...
Identifiers
Publication date
1997-03
Defense date
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This paper presents a wider characterization of memory in time series and of co integration in terms of information-theoretic statistics such as the entropy and the mutual information between pairs of variables. This suggests a new methodology for exploratory data analysis and for testing the hypothesis of long-memory and of the existence of a co integrating relationship. We illustrate the performances of the new techniques with some simulation experiments, and finally apply them to the analysis of the relationship between pairs of financial time series from a foreign exchangerate and a stock return markets.
Description
Keywords
Information-theoretic statistics, long memory, cointegration, nonlinearity, causality