Survey of lakes or streams to assess and monitor their environmental conditions are often conducted using complex designs. Inference on descriptive parameters - as population totals or means of variables of interest - that accounts for the sampling procedures can be borrowed from design based survey theory. In particular, design based inference assisted by superpopulation models have proved to be a useful tool to employ auxiliary information available on the finite population of interest. Generalized regression estimation (Sarndal, Swensson and Wretman, 1992) and calibration estimation (Deville and Sarndal, 1992) employ auxiliary information in the form of population totals or means of auxiliary variables and are now widely used in many National Institutes of Statistics to produce a wide range of official statistics. In their original formulation, these methodologies assume - either explicitly or implicitly - a linear superpopulation model for the relationship between a variable of interests and the auxiliary ones. However, they both provide very flexible frameworks that have been recently exploited in literature to extend the class of superpopulation assisting models, when auxiliary information is complete, i.e. when the value taken by the auxiliary variable is known for each unit in the population. In particular, we will review and discuss the possibility of assuming non parametric regression models that allow more flexible modeling of the variables of interest (Breidt and Opsomer, 2007; Montanari and Ranalli, 2005). This has particular relevance for inference on finite populations of lakes or streams, in which the variables of interest may have very complicated dependencies on geographically referenced covariates. Auxiliary information in these contexts comes from remote sensing and GIS models and may be particularly rich.

Non parametric regression model-assisted estimation for finite environmental populations

RANALLI, Maria Giovanna
2009

Abstract

Survey of lakes or streams to assess and monitor their environmental conditions are often conducted using complex designs. Inference on descriptive parameters - as population totals or means of variables of interest - that accounts for the sampling procedures can be borrowed from design based survey theory. In particular, design based inference assisted by superpopulation models have proved to be a useful tool to employ auxiliary information available on the finite population of interest. Generalized regression estimation (Sarndal, Swensson and Wretman, 1992) and calibration estimation (Deville and Sarndal, 1992) employ auxiliary information in the form of population totals or means of auxiliary variables and are now widely used in many National Institutes of Statistics to produce a wide range of official statistics. In their original formulation, these methodologies assume - either explicitly or implicitly - a linear superpopulation model for the relationship between a variable of interests and the auxiliary ones. However, they both provide very flexible frameworks that have been recently exploited in literature to extend the class of superpopulation assisting models, when auxiliary information is complete, i.e. when the value taken by the auxiliary variable is known for each unit in the population. In particular, we will review and discuss the possibility of assuming non parametric regression models that allow more flexible modeling of the variables of interest (Breidt and Opsomer, 2007; Montanari and Ranalli, 2005). This has particular relevance for inference on finite populations of lakes or streams, in which the variables of interest may have very complicated dependencies on geographically referenced covariates. Auxiliary information in these contexts comes from remote sensing and GIS models and may be particularly rich.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11391/1030085
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact