1
$\begingroup$

Dudley's chaining integral is commonly used to bound Rademacher complexities. I recall seeing several papers give this as the reference

@ARTICLE{MR512411,
  author = {Dudley, R. M.},
  title = {Central limit theorems for empirical measures},
  journal = {Ann. Probab.},
  year = {1978},
  volume = {6},
  pages = {899--929 (1979)},
  number = {6},
  coden = {APBYAE},
  fjournal = {The Annals of Probability},
  issn = {0091-1798},
  mrclass = {60F05 (28C20 60B10 60F17)},
  mrnumber = {MR512411 (81k:60029a)},
  mrreviewer = {P. R{\'e}v{\'e}sz}
}

but I don't think the result in question actually appears in that paper. Could anyone point me to the definitive reference?

$\endgroup$
3
$\begingroup$

I think the proper reference is:

@article{Dudley1967290,
    Author = {R.M Dudley},
    Doi = {http://dx.doi.org/10.1016/0022-1236(67)90017-1},
    Issn = {0022-1236},
    Journal = {Journal of Functional Analysis},
    Number = {3},
    Pages = {290 - 330},
    Title = {The sizes of compact subsets of Hilbert space and continuity of Gaussian processes},
    Url = {http://www.sciencedirect.com/science/article/pii/0022123667900171},
    Volume = {1},
    Year = {1967},
    Bdsk-Url-1 = {http://www.sciencedirect.com/science/article/pii/0022123667900171},
    Bdsk-Url-2 = {http://dx.doi.org/10.1016/0022-1236(67)90017-1}}

See Theorem 3.1. for the relevant result, although stated in terms of continuity of Gaussian processes.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.