ruth@utstat.uucp (Ruth Croxford) (04/01/89)
Colloquium Series, Department of Statistics, University of Toronto Topic: Jackknife Variance Estimation Speaker: Jun Shao, University of Waterloo Date: Friday, April 7, 1989 3:00 p.m. Place: Room 2110, Sidney Smith Hall, 100 St. George St., University of Toronto Abstract: The traditional delete-1 jackknife variance estimator is asymtotically correct for smooth point estimators but is inconsistent for non-smooth estimators. This deficiency can be rectified by using a more general jackknife with d, the number of observations deleted, depending on a smoothness measure of the point estimator. Our general theory explains why and when jackknife works or fails. It also shows that for non-smooth estimators such as the sample quantiles, d has to go to infinity at a certain rate. Two methods for efficient computation of the delete-d jackknife variance estimators are studied. The robustness of the jack- knife against heavy tails of the population and error heteroscedasticity is also discussed. ------------ Coffee and tea will be served in the De Lury Lounge (SS6006) at 2:30 p.m.