Timeline for Linear model where the data has uncertainty, using R
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 19, 2016 at 12:26 | vote | accept | Gimelist | ||
| Sep 19, 2016 at 12:15 | answer | added | jwimberley | timeline score: 16 | |
| Sep 19, 2016 at 11:58 | comment | added | jwimberley | See also this question here:stats.stackexchange.com/questions/113987/… | |
| Sep 19, 2016 at 9:30 | history | edited | Gimelist | CC BY-SA 3.0 |
added 1459 characters in body
|
| Sep 19, 2016 at 8:49 | comment | added | Pascal |
lm will use the normalized variances as weights and then assume that your model is statistically valid to estimate parameters uncertainty. If you think that this is not the case (error bars too small or too large), you should not trust any uncertainty estimate.
|
|
| Sep 19, 2016 at 8:33 | comment | added | Ferdi |
If you know the distribution of the data you can bootstrap it using the boot package in R. Afterwards you can let a linear regression run over the bootstrapped data set.
|
|
| Sep 19, 2016 at 8:04 | history | asked | Gimelist | CC BY-SA 3.0 |