 Conditional entropy and randomness in financial time seriesM. D. London, A. K. Evans and M. J. Turner Quantitative Finance, 2001, vol. 1, issue 4, 414-426 Abstract: This paper investigates the fundamental question of whether or not financial time series can be predicted. It does so using the conditional entropy measure commonly used in information theory and statistical physics. This approach is appropriate because it will reveal patterns of arbitrary complexity. That is, this method will not just reveal linear correlation, or any specific nonlinear correlation, it will reveal any patterns in the data. Interesting discoveries include the fact that there is a degree of consistency amongst data sets and the ability to clearly distinguish between stock market indices and exchange rate time series. The fact that the magnitude of price movements is more correlated than the direction is verified and quantified in the context of daily data. The main result is that above certain time scales financial time series are random, but below this threshold there are situations where knowing a portion of the past can bring a statistically significant amount of certainty about the future. Date: 2001 Downloads: (external link) Related works: Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text Persistent link: https://EconPapers.repec.org/RePEc:taf:quantf:v:1:y:2001:i:4:p:414-426 Ordering information: This journal article can be ordered from DOI: 10.1088/1469-7688/1/4/302 Access Statistics for this article Quantitative Finance is currently edited by Michael Dempster and Jim Gatheral More articles in Quantitative Finance from Taylor & Francis Journals |
Is your work missing from RePEc? Here is how to contribute.
Questions or problems? Check the EconPapers FAQ or send mail to .
EconPapers is hosted by the
School of Business at Örebro University.