This paper introduces an ordinary kriging predictor for histogram data. We assume that the input data is a set of histograms which summarize data observed in a geographic area. Our aim is to predict the histogram of data in a spatial location where it is not possible to get records. We consider the histograms as random elements of a L2-Wasserstein space. The isometry from the Wasserstein space of probability measures to the subset of L2(0,1) of quantile functions, allows us to introduce a linear predictor which uses the quantile functions associated with the histograms.
A new ordinary kriging predictor for histogram data in L2-Wasserstein space
A. Balzanella
Writing – Original Draft Preparation
;R. VerdeWriting – Original Draft Preparation
;A. IrpinoWriting – Original Draft Preparation
2019
Abstract
This paper introduces an ordinary kriging predictor for histogram data. We assume that the input data is a set of histograms which summarize data observed in a geographic area. Our aim is to predict the histogram of data in a spatial location where it is not possible to get records. We consider the histograms as random elements of a L2-Wasserstein space. The isometry from the Wasserstein space of probability measures to the subset of L2(0,1) of quantile functions, allows us to introduce a linear predictor which uses the quantile functions associated with the histograms.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.
