Answers are grounded in this paper and cite their sources.
By applying the Chebyshev inequality directly to a linear discriminant function, Yau and Lin [1] obtained a distribution-free upper bound to the probability of error by identifying two classes. This letter provides, in a similar way, another upper bound of the same applicability which is sharper when good class separation exists.
No takes yet. Share an insight, caveat, or question.
W.J.J. Rey (1976) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: