Asymptotic linear representation
= Asymptotic linear representation
An <estimator> has an asymptotic linear representation with <influence function> $\psi$ when
$$
\sqrt n(\widehat\theta-\theta)
=\frac1{\sqrt n}\sum_{i=1}^n\psi(X_i)+o_p(1).
$$
= Asymptotic linear representation
An <estimator> has an asymptotic linear representation with <influence function> $\psi$ when
$$
\sqrt n(\widehat\theta-\theta)
=\frac1{\sqrt n}\sum_{i=1}^n\psi(X_i)+o_p(1).
$$