Skip to main content
10 events
when toggle format what by license comment
Nov 20, 2022 at 10:11 comment added Hermi @ThomasKojar Sorry, if possible, would you mind taking a look at this question mathoverflow.net/questions/434736/…? Thank you!
Nov 5, 2022 at 0:30 comment added Thomas Kojar yes indeed. But still I advise checking out their book if possible, they go into depth with many formulas.
Nov 5, 2022 at 0:28 comment added Thomas Kojar thank you. Indeed, in that book the result is "iff" with continuity for the distribution. Even for atomic distributions, they have nice formulas in terms of t=0.
Nov 4, 2022 at 23:42 history edited Thomas Kojar CC BY-SA 4.0
added 11 characters in body
Nov 4, 2022 at 23:08 comment added dohmatob Thanks for the update. Much clearer now. I see how your core argument goes through, but there are still a few of issues. First, "continuity in $t$-variable" is not for free. This will only hold if, for every $w$, the random variable $X'w$ has density. For this to hold, it is sufficient to assume that the random vector $X$ has density. Typo: Given the definition of $S_t$, I don't know what you mean by "$x_n \in S_t$". I have also posted an answer below which only assumes $X$ has density, and then uses a truncation argument (maybe this is what you had in mind in your very first comment ?).
Nov 4, 2022 at 19:39 history edited Thomas Kojar CC BY-SA 4.0
edited body
Nov 4, 2022 at 19:17 history edited Thomas Kojar CC BY-SA 4.0
added 561 characters in body
Nov 4, 2022 at 7:44 comment added dohmatob Thanks for the input (upvoted). Unfortunately, the linked book is not accessible. In particular, I don't have access to the statement of Theorem 1.7.4 (which you referenced), and so I can't tell if it provides any insight on the problem at hand.
Nov 4, 2022 at 3:39 history edited Thomas Kojar CC BY-SA 4.0
added 162 characters in body
Nov 3, 2022 at 23:30 history answered Thomas Kojar CC BY-SA 4.0