Bernstein's inequality

GPTKB entity

Statements (18)
Predicate Object
gptkbp:instanceOf gptkb:inequality
gptkbp:appliesTo sum of independent random variables
gptkbp:field gptkb:probability_theory
mathematical analysis
https://www.w3.org/2000/01/rdf-schema#label Bernstein's inequality
gptkbp:namedAfter gptkb:Sergei_Bernstein
gptkbp:provides upper bound on probability of large deviations
gptkbp:publishedIn 1924
gptkbp:relatedTo gptkb:Hoeffding's_inequality
gptkb:Azuma's_inequality
gptkb:Chernoff_bound
gptkbp:sentence For independent zero-mean random variables bounded by M, the probability that their sum exceeds t is bounded by exp(-t^2/(2σ^2 + 2Mt/3)).
gptkbp:type concentration inequality
gptkbp:usedIn gptkb:machine_learning
gptkb:empirical_process_theory
statistics
gptkbp:bfsParent gptkb:Joseph_Bernstein
gptkbp:bfsLayer 5