Timeline for Confidence interval for the standard deviation on a bimodal distribution
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 21, 2014 at 10:44 | comment | added | Glen_b | Also see here | |
| Jul 21, 2014 at 9:33 | comment | added | Glen_b | (ctd) ... or see some of the details here. Finally, you might try simulation or bootstrapping. | |
| Jul 21, 2014 at 9:32 | comment | added | Glen_b | There are several possible options. If you model the bimodal distribution (say as a mixture of unimodal distributions), you might be able to derive results for inference on the variance or standard deviation. If the samples are large enough, you can still apply the result of van der Vaart, if you know the population mean and variance of the sample variance, or if the samples are large enough that you can act as if Slutsky's theorem was to "kick in"...(ctd) | |
| Jul 21, 2014 at 6:45 | comment | added | user52407 | Thank you. So how would you make inference on the standard deviation given a bimodal distribution like the one in the image? I can get two standard deviations out of it, one for each distribution within the bimodal, but is there a way to pool them together and get a confidence interval ? | |
| Jul 20, 2014 at 20:52 | history | edited | Glen_b | CC BY-SA 3.0 |
added 585 characters in body
|
| Jul 20, 2014 at 20:34 | history | edited | Glen_b | CC BY-SA 3.0 |
added 205 characters in body
|
| Jul 20, 2014 at 20:22 | history | edited | Glen_b | CC BY-SA 3.0 |
added 294 characters in body
|
| Jul 20, 2014 at 20:13 | history | edited | Glen_b | CC BY-SA 3.0 |
added 8 characters in body
|
| Jul 20, 2014 at 20:08 | history | answered | Glen_b | CC BY-SA 3.0 |