Universal probability-free prediction
Machine Learning
2017-04-05 v2
Abstract
We construct universal prediction systems in the spirit of Popper's falsifiability and Kolmogorov complexity and randomness. These prediction systems do not depend on any statistical assumptions (but under the IID assumption they dominate, to within the usual accuracy, conformal prediction). Our constructions give rise to a theory of algorithmic complexity and randomness of time containing analogues of several notions and results of the classical theory of Kolmogorov complexity and randomness.
Keywords
Cite
@article{arxiv.1603.04283,
title = {Universal probability-free prediction},
author = {Vladimir Vovk and Dusko Pavlovic},
journal= {arXiv preprint arXiv:1603.04283},
year = {2017}
}
Comments
27 pages