New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Merged by Bors] - feat: add last implication of portmanteau characterizations of weak convergence #8097
[Merged by Bors] - feat: add last implication of portmanteau characterizations of weak convergence #8097
Conversation
…onst and 7 friends.
…ons (except with subtraction counterparts have to assume AddCommGroup and one additional covariance property).
…StronglyMeasurable?).
…ions to one file. Generalize to usual typeclasses.
…r AntitoneOn functions?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
bors d+
Thanks!
✌️ kkytola can now approve this pull request. To approve and merge a pull request, simply reply with |
Co-authored-by: sgouezel <sebastien.gouezel@univ-rennes1.fr>
bors r+ |
…onvergence (#8097) This PR adds the last missing implication of the general case of portmanteau equivalent characterizations of convergence in distribution: a sufficient condition for convergence in distribution of a sequence of probability measures is that for all open sets the candidate limit measure is at most the liminf of the measures. Co-authored-by: Kalle <kalle.kytola@aalto.fi> Co-authored-by: kkytola <39528102+kkytola@users.noreply.github.com>
Build failed: |
bors r+ |
…onvergence (#8097) This PR adds the last missing implication of the general case of portmanteau equivalent characterizations of convergence in distribution: a sufficient condition for convergence in distribution of a sequence of probability measures is that for all open sets the candidate limit measure is at most the liminf of the measures. Co-authored-by: Kalle <kalle.kytola@aalto.fi> Co-authored-by: kkytola <39528102+kkytola@users.noreply.github.com>
Pull request successfully merged into master. Build succeeded: |
…onvergence (#8097) This PR adds the last missing implication of the general case of portmanteau equivalent characterizations of convergence in distribution: a sufficient condition for convergence in distribution of a sequence of probability measures is that for all open sets the candidate limit measure is at most the liminf of the measures. Co-authored-by: Kalle <kalle.kytola@aalto.fi> Co-authored-by: kkytola <39528102+kkytola@users.noreply.github.com>
…onvergence (#8097) This PR adds the last missing implication of the general case of portmanteau equivalent characterizations of convergence in distribution: a sufficient condition for convergence in distribution of a sequence of probability measures is that for all open sets the candidate limit measure is at most the liminf of the measures. Co-authored-by: Kalle <kalle.kytola@aalto.fi> Co-authored-by: kkytola <39528102+kkytola@users.noreply.github.com>
In my earlier PR #8097 there there was a typo `variable {Ω : Type}` which should have been a universe polymorphic `variable {Ω : Type*}`. This PR fixes 1 character. Apologies!
…onvergence (#8097) This PR adds the last missing implication of the general case of portmanteau equivalent characterizations of convergence in distribution: a sufficient condition for convergence in distribution of a sequence of probability measures is that for all open sets the candidate limit measure is at most the liminf of the measures. Co-authored-by: Kalle <kalle.kytola@aalto.fi> Co-authored-by: kkytola <39528102+kkytola@users.noreply.github.com>
In my earlier PR #8097 there there was a typo `variable {Ω : Type}` which should have been a universe polymorphic `variable {Ω : Type*}`. This PR fixes 1 character. Apologies!
This PR adds the last missing implication of the general case of portmanteau equivalent characterizations of convergence in distribution: a sufficient condition for convergence in distribution of a sequence of probability measures is that for all open sets the candidate limit measure is at most the liminf of the measures.
(There are more equivalent conditions dubbed portmanteau for the special case of convergence in distribution on R, though. Some of them are easy and I'll add soon. The ones with characteristic functions require characteristic functions.)
One restriction in this implication is that it really talks about a sequence and the
atTop
filter rather than a general filter like the other implications. The reason is that Fatou's lemma is used (and it is not obvious to me that it could be generalized to much more than countably generated filters in countable types).One design choice in the proof is that I could not use the relevant lemma
Monotone.map_liminf_of_continuousAt
to commuteliminf
withENNReal.toReal
, since the lemma assumes monotonicity, butENNReal.toReal
fails to be monotone because of +infty. I used a truncated variantENNReal.truncateToReal
instead. Perhaps a more principled strategy would have been to prove a generalizationMonotoneOn.map_liminf_of_continuousAt
, but it was less straightforward to do the correct statement and proof.