Abstract
Cokriging is a geostatistical technique that is used for spatial prediction when realizations of a random field are available. If a secondary variable is cross-correlated with the primary variable, both variables may be employed for prediction by means of cokriging. In this work, we propose a predictive model that is based on cokriging when the secondary variable is functional. As in the ordinary cokriging, a co-regionalized linear model is needed in order to estimate the corresponding auto-correlations and cross-correlations. The proposed model is utilized for predicting the environmental pollution of particulate matter when considering wind speed curves as functional secondary variable.
Subject
General Mathematics,Engineering (miscellaneous),Computer Science (miscellaneous)
Reference51 articles.
1. Model-Based Geoestatistics;Diggle,2007
2. Statistics for Spatial Data;Cressie,1993
3. Constructing and fitting models for cokriging and multivariable spatial prediction
4. Geostatistics: Modeling Spatial Uncertainty;Chiles,1999
5. Air contaminant statistical distributions with application to PM10 in Santiago, Chile;Marchant;Rev. Environ. Contam. Toxicol.,2013
Cited by
29 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献