Multivariate density estimation for interval‐censored data with application to a forest fire modelling problem. Issue 6 (15th June 2016)

• Record Type: Journal Article
• Authors: Braun, W. J.; Stafford, J. E.
• Is Part Of: Environmetrics. Volume 27:Issue 6(2016:Sep.)
• Page Start: 345
• Page End: 354
• Publication Date: 2016-06-15
• Subjects: density estimation -- interval‐censoring -- cross‐validation -- EMS, local likelihood -- Duff Moisture Code -- holdover; Environmental sciences -- Statistical methods -- Periodicals; 550.72
• DOI: 10.1002/env.2396
• Languages: English
• ISSNs: 1180-4009

Abstract

Abstract : Local likelihood‐based density estimation methods are developed for multivariate interval‐censored data. An extension of the Nadaraya–Watson local regression estimator to the interval‐censored data context arises in a natural way. The conditional density estimation scheme is used to study the holdover distribution of lightning‐caused wildfire ignitions in Northern Ontario. Copyright © 2016 John Wiley & Sons, Ltd.