Estimating a Signal with Noisy Nuisance Parameters
Ya’acov Ritov · Biometrika · 1989
We describe two models in which n records of a signal in white noise are taken. In the first model the signal parameters of interest do not change between records, but the amplitude varies in a random way. In the second model, the location is the random nuisance parameter. We describe an efficient estimator for the first model. For the second model an estimator is described which is n½ consistent, invariant to the distribution of the location and efficient for a particular distribution.