Minimax deviation strategies for machine learning and recognition with short learning samples

TL;DR

This paper introduces minimax deviation learning as a novel approach to improve machine learning and recognition tasks when only small learning samples are available, addressing flaws in traditional methods.

Contribution

It proposes the concept of minimax deviation learning, which overcomes limitations of maximum likelihood and minimax learning in small sample scenarios.

Findings

Minimax deviation learning is free of flaws found in traditional methods.

The approach improves recognition accuracy with small samples.

The paper provides theoretical foundations for the new learning strategy.

Abstract

The article is devoted to the problem of small learning samples in machine learning. The flaws of maximum likelihood learning and minimax learning are looked into and the concept of minimax deviation learning is introduced that is free of those flaws.