Thomas Lumley • 4/30/2019

Local asymptotic minimax, and nearly-true models

The article discusses the comparison of estimators (specifically AIPW and semiparametric-efficient estimators) in regression models with two-phase sampling when the assumed outcome model is 'nearly true.' It delves into theoretical concepts like contiguity, local asymptotic minimax theorems, and mean squared error under sequences of models that are statistically indistinguishable from the assumed model at given sample sizes.

0 comments

#Regression Models #Statistical Inference #Asymptotic Theory