Threshold of overflow probability in terms of smooth max-entropy for variable-length compression allowing errors
Threshold of overflow probability in terms of smooth max-entropy for variable-length compression allowing errors
复制标题
允许错误的可变长度压缩的平滑最大熵的溢出概率阈值
DOI：
--
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
T. Matsushima
中科院分区：
文献类型：
--
作者：
Shota Saito;T. Matsushima
本文讨论允许错误概率的一次性固定长度到可变长度源编码。我们采用溢出概率的准则，对前缀码和非前缀码进行处理。在错误概率和溢出概率以正常数为界的条件下，研究了溢出概率阈值的下确界。我们证明这个阈值是根据平滑最大熵来评估的，并阐明了前缀和非前缀代码之间阈值的差异。此外，我们阐明了与无损编码相比允许非零错误的好处。
This paper deals with one-shot fixed-to-variable length source coding allowing error probability. We adopt the criterion of the overflow probability and treat prefix and non-prefix codes. The infimum of the threshold of the overflow probability is investigated under the condition that the error probability and the overflow probability are bounded above by positive constants. We show this threshold is evaluated in terms of the smooth max-entropy and clarify the difference of the thresholds between prefix and non-prefix codes. Further, we elucidate the benefit of allowing a nonzero error compared with lossless coding.