首页    期刊浏览 2024年11月24日 星期日
登录注册

文章基本信息

  • 标题:Entropy and Higher Moments of Information
  • 本地全文:下载
  • 作者:Helmut Jürgensen (University of Waterloo ; Canada) David E. Matthews (University of Waterloo, Canada)
  • 期刊名称:Journal of Universal Computer Science
  • 印刷版ISSN:0948-6968
  • 出版年度:2010
  • 卷号:16
  • 期号:5
  • 出版社:Graz University of Technology and Know-Center
  • 摘要:

    Abstract: The entropy of a finite probability space or, equivalently, a memoryless source is the average information content of an event. The fact that entropy is an expectation suggests that it could be quite important in certain applications to take into account higher moments of information and parameters derived from these like the variance or skewness. In this paper we initiate a study of the higher moments of information for sources without memory and sources with memory. We derive properties of these moments for information defined in the sense of Shannon and indicate how these considerations can be extended to include the concepts of information in the sense of Aczél or Rényi. For memoryless sources, these concepts are immediately supported by the usual definitions of moments; for general stationary sources, let alone general sources, no such applicable framework seems to exist; on the other hand, the special properties of stationary Markov sources suggest such definitions which are both, well-motivated and mathematically meaningful.

  • 关键词:entropy, information, moments of information, variance of information
国家哲学社会科学文献中心版权所有