arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2304.10752cs.ITcs.AImath.IT
收藏

算法信息可预测性

Algorithmic Information Forecastability

Glauco Amigo, Daniel Andrés Díaz-Pachón, Robert J. Marks, Charles Baylis

AI总结:

本文利用算法信息论定义仅依赖数据本身的可预测性度量,并提出预言、精确和概率三类可预测性,为时间序列预测与分类提供了统一的理论框架。

摘要:

并非所有时间序列的结果都能被预测,例如抛掷一枚公平硬币的结果。而另一些序列,如重复的{01}序列{010101...},则可以被精确预测。算法信息论可以提供一种介于这两个极端之间的可预测性度量。可预测性的程度仅取决于数据本身。对于带标签数据的预测(或分类),我们提出三类可预测性:oracle forecastability(预言可预测性),指预测始终精确;precise forecastability(精确可预测性),指误差有界;以及probabilistic forecastability(概率可预测性),指其他任意预测。每种情况都给出了示例。

英文摘要:

The outcome of all time series cannot be forecast, e.g. the flipping of a fair coin. Others, like the repeated {01} sequence {010101...} can be forecast exactly. Algorithmic information theory can provide a measure of forecastability that lies between these extremes. The degree of forecastability is a function of only the data. For prediction (or classification) of labeled data, we propose three categories for forecastability: oracle forecastability for predictions that are always exact, precise forecastability for errors up to a bound, and probabilistic forecastability for any other predictions. Examples are given in each case.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑