Probability tails of Gaussian extrema (Q805058): Difference between revisions

A linear map L from the Hilbert space H into the real Gaussian variables with E Lx\(=0\), E LxLy\(=(x,y)\) is called isonormal Gaussian process. Inequalities on the tail of the distribution \(P\{\sup_{x\in {\mathcal C}}Lx>\lambda \}\) (where \({\mathcal C}\) is a suitable subset of H) are given. The results are used to prove similar inequalities for set-indexed Brownian sheet and Brownian bridge.