Timeline for Calculate accelerated bootstrap interval in R
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 23, 2019 at 15:30 | comment | added | Sal Mangiafico | See also this StackExchange link for a couple of additional references. | |
| Nov 23, 2019 at 15:28 | comment | added | Sal Mangiafico |
Thanks. The documentation for the boot package cites "Chapter 5 of Davison and Hinkley (1997)" for the procedures for the confidence intervals. So, that's probably the definitive source for questions on the procedures there. The documentation describes the bca option as "adjusted bootstrap percentile" even though most places on the internet say it is "bias correction with acceleration" (or something similar).
|
|
| Nov 23, 2019 at 5:10 | comment | added | knrumsey | +1, very cool! I would almost guarantee the methods ‘perc’ and ‘bca’ are the same as above. (BCa stands for bias correction with acceleration constant). The slight discrepancies can be do to MC error (even though you set the seed, the Bootstrap procedure might be slightly different). The other reason could be a slight difference in the estimation procedure of the acceleration constant. | |
| Nov 23, 2019 at 3:47 | history | edited | Sal Mangiafico | CC BY-SA 4.0 |
deleted 58 characters in body
|
| Nov 23, 2019 at 1:51 | history | edited | Sal Mangiafico | CC BY-SA 4.0 |
deleted 335 characters in body
|
| Nov 23, 2019 at 1:43 | history | answered | Sal Mangiafico | CC BY-SA 4.0 |