Hazard rate estimation on random fields
Consider observations (representing lifelengths) taken on a random field indexed by lattice points. Our purpose is to estimate the hazard rate r(x), which is the rate of failure at time x for the survivors up to time x. We estimate r(x) by the nonparametric estimator constructed in terms of a kernel-type estimator for f(x) and the natural estimator for. Under some general mixing assumptions, the limiting distribution of the estimator at multiple points is shown to be multivariate normal. The result is useful in establishing confidence bands for r(x) with x in an interval.
Year of publication: |
2007
|
---|---|
Authors: | Li, Jiexiang ; Tran, Lanh Tat |
Published in: |
Journal of Multivariate Analysis. - Elsevier, ISSN 0047-259X. - Vol. 98.2007, 7, p. 1337-1355
|
Publisher: |
Elsevier |
Keywords: | Random field Spatial process Hazard rate Kernel density estimator Asymptotic normality |
Saved in:
Saved in favorites
Similar items by person
-
Order statistics for nonstationary time series
Tran, Lanh Tat, (1993)
-
Kernel density estimation under dependence
Tran, Lanh Tat, (1990)
-
Kernel density estimation for random fields (density estimation for random fields)
Carbon, Michel, (1997)
- More ...