Let Zn = {(x1, y1),...,(xn, yn)}, be a sample for the logistic model, where the first coordinate

Question:

Let Zn = {(x1, y1),...,(xn, yn)}, be a sample for the logistic model, where the first coordinate of each xi is 1 if the model contains an intercept. Consider a new sample Z∗

n = {(−x1, 1 − y1),...,(−xn, 1 − yn)}.

(a) Explain why is desirable that an estimator ????̂ satisfies the equivariance property ????̂(Z∗

n ) = ????̂(Zn).

(b) Show that M-, WM- and CUBIF estimators satisfy this property.

Fantastic news! We've Found the answer you've been seeking!

Step by Step Answer:

Related Book For  book-img-for-question

Robust Statistics Theory And Methods

ISBN: 9781119214687

2nd Edition

Authors: Ricardo A. Maronna, R. Douglas Martin, Victor J. Yohai, Matías Salibián-Barrera

Question Posted: