We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.IT

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Computer Science > Information Theory

Title: A rapid algorithm to calculate joint probability matrices for joint entropies of arbitrary order

Abstract: There is no closed form analytical equation or quick method to calculate probabilities based only on the entropy of a signal or process. Except in the cases where there are constraints on the state probabilities, one must typically derive the underlying probabilities through search algorithms. These become more computationally expensive as entropies of higher orders are investigated. In this paper, a method to calculate a joint probability matrix based on the entropy for any order is elaborated. With this method, only first order entropies need to be successfully calculated while the others are derived via multiplicative cascades.
Comments: 4 pages 1 figure
Subjects: Information Theory (cs.IT)
Cite as: arXiv:1501.05606 [cs.IT]
  (or arXiv:1501.05606v1 [cs.IT] for this version)

Submission history

From: Reginald Smith [view email]
[v1] Sun, 30 Nov 2014 20:59:32 GMT (236kb)

Link back to: arXiv, form interface, contact.