...
首页> 外文期刊>South African statistical journal >A NOTE ON EXTENDED ARIMOTO'S ENTROPIES
【24h】

A NOTE ON EXTENDED ARIMOTO'S ENTROPIES

机译:关于扩展ARIMOTO的熵的注释

获取原文
获取原文并翻译 | 示例
           

摘要

Arimoto (1971) introduced, among other things, a class of entropies for probability distributions on any finite set of elements, which includes Shannon's entropy (1948) as a special case. Restricted to the set y~2 of probability distributions P = (t, 1 - t) on a set with only two elements his class of entropies is given in terms of h_a(t)={(1/(1-α)[1-t~(1/α)+(1-t)~(1/α))~α] if α ∈ (0,∞){1}) (-[t lnt + (1-t)ln(1-t)] if α=1) (min(t,1-t) if α=0) As Vajda (2009) extended a certain class of Csiszar's /-divergences, which are closely related to Arimoto's entropies to all parameters a s α∈R, the authors of this note generalised the special case of Arimoto's entropies for probability distributions P∈ y)_2 to all α∈R (De Wet and Osterreicher, 2016). It turns out that these entropies are given for negative α=-k, k ∈ (0,∞), by h_(-k)(t)=1/(1+k) (t(1-t))/([t~(1/k)+(1-t)~(1/k)]k),t∈[0,1]. In the present note their extension to probability distributions P ∈ y_n for n ≥ 2 is investigated. In addition, a comparison of Arimoto's extended class of entropies with Renyi's and Tsallis' classes, is given. For the axiomatic characterization of the latter two classes of entropies we refer to the survey paper by Csiszar (2008).
机译:除其他事项外,Arimoto(1971)引入了一类用于任何有限元素集上概率分布的熵,其中包括香农的熵(1948)作为特例。限制在只有两个元素的集合上的概率分布的集合y〜2 P =(t,1-t),他的熵类别以h_a(t)= {(1 /(1-α)[如果α∈(0,∞){1})(-[t lnt +(1-t)ln(1-t〜(1 /α)+(1-t)〜(1 /α))〜α] 1-t)]如果α= 1)(如果α= 0则为min(t,1-t)随着Vajda(2009)扩展了一类Csiszar的/-散度,与所有参数的Arimoto熵密切相关,如α∈R,该注释的作者将Arimoto熵的特殊情况推广到了所有α∈R的概率分布P∈y)_2(De Wet and Osterreicher,2016)。事实证明,这些熵是针对负α= -k,k∈(0,∞)通过h _(-k)(t)= 1 /(1 + k)(t(1-t))/( [t〜(1 / k)+(1-t)〜(1 / k)] k),t∈[0,1]。在本说明中,研究了当n≥2时它们对概率分布P∈y_n的扩展。另外,给出了有本的熵的扩展类与人意类和Tsallis类的比较。对于后两类熵的公理化特征,请参考Csiszar(2008)的调查论文。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号