Timeline for Constants in the Rosenthal inequality
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 28, 2022 at 12:48 | comment | added | Iosif Pinelis | @ThomasDybdahlAhle : Yes, these constant factors are achievable -- when the $X_i$'s are iid symmetric with a $3$-point support set; see Lemma 6.5 in the first linked paper. | |
| Aug 28, 2022 at 12:37 | history | edited | Iosif Pinelis | CC BY-SA 4.0 |
added 231 characters in body
|
| Aug 28, 2022 at 11:32 | comment | added | Thomas Dybdahl Ahle | do you have an intuition (or theorem) for when $C_2=(c\sqrt t)^t$ and $C_1=(ct/\log t)^t$ is achievable? Like I suppose it is for simple binomial and poisson distributions. Or maybe that's not even true? | |
| Aug 25, 2022 at 13:30 | comment | added | Iosif Pinelis | @ThomasDybdahlAhle : Yes, this is true. | |
| Aug 25, 2022 at 9:00 | comment | added | Thomas Dybdahl Ahle | So you can get $C_2(t) = (c t^{1/2+\varepsilon})^t$ and $C_1(t) = \left(\frac{ct}{\log t}\right)^t$ for any $\varepsilon>0$? | |
| Apr 23, 2018 at 19:03 | history | answered | Iosif Pinelis | CC BY-SA 3.0 |