Premium
Using inter‐event functions for bandwidth selection in disease incidence ratio estimation
Author(s) -
Chadæuf J.,
Lebond A.,
Senoussi R.
Publication year - 2004
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.685
Subject(s) - estimator , bandwidth (computing) , bivariate analysis , statistics , mathematics , kernel density estimation , univariate , parametric statistics , selection principle , selection (genetic algorithm) , econometrics , computer science , multivariate statistics , artificial intelligence , telecommunications , reliability engineering , engineering
We propose a distance based criterion to choose the optimal bandwidth for a non‐parametric kernel estimator of the mark probability of a non‐stationary bivariate location process. The selected bandwidth achieves the minimum standardized distance between the observed inter‐event distance distribution and its expectation under the estimated model. This selection principle is performed to map the intensity of the incidence ratio of a West‐Nile outbreak in the South of France. Copyright © 2004 John Wiley & Sons, Ltd.