Timeline for The theory behind the weights argument in R when using lm()
Current License: CC BY-SA 3.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 20, 2016 at 2:29 | vote | accept | Clarinetist | ||
| Nov 20, 2016 at 2:21 | comment | added | mark999 |
Yes, although the weights only have to be proportional to 1/variance, not necessarily equal. For example, if you use weights <- c(50, 85, 75)/2 in your example, you get the same result.
|
|
| Nov 20, 2016 at 2:14 | comment | added | Clarinetist | Thank you for clearing up the incorrect design matrix, especially! I'm quite rusty on this material. So, as one last question, does this mean that $\text{Var}(\boldsymbol\epsilon) = \text{diag}(1/\text{weights})$ in the WLS assumptions? | |
| Nov 20, 2016 at 2:11 | history | answered | mark999 | CC BY-SA 3.0 |