I have two samples of data each estimates of a position x, y with Gaussian noise. One source has a larger variance than the other. Is this source in any way useful in providing a better estimate of the true position than the source with lower variance or should it just be discounted?
Asked
Active
Viewed 41 times
0
nickponline
- 546
- 4
- 11
-
Do you know thwe variances -- or the ratio of the variances? – Russ Lenth Aug 21 '14 at 18:44
-
More information is also available at http://stats.stackexchange.com/questions/5392 and http://stats.stackexchange.com/questions/8502 (which solves a generalization of this problem: the estimator is given by regressing the data against a constant). – whuber Aug 21 '14 at 18:56
-
@Russ Lenth Yes they are known. – nickponline Aug 21 '14 at 20:33