You are here: Wiki>AI_GEOSTATS Web>AI_GEOSTATSPapers>Papers20100623105619 (13 Aug 2010, theresiafreska)Edit Attach
Title: A Methodological Study of the Application of the Maximum Entropy Estimator to Spatial Interpolation

Date: 1 May 1998

Authors: Yuh-Ming Lee

Link: LeeYM.pdf


This study applies the maximum entropy estimator to the interpolation of daily rainfall measurements in Switzerland taken on April 26, 1986. The major purpose of the paper is to introduce the methodology of maximum-entropy spatial interpolation. The accuracy of the estimation measured by the coefficient of determination is therefore only at the value of 0.0114. Moreover, the estimation errors are, to some extent, correlated to the real measurements. The larger the measurement is, the more severe the error becomes. The structure analysis on the 100 measurements used for estimation depicts the properties of anisotropy and non-stationarity in spite of the assumptions of second-order stationarity and isotropic correlation. The estimation errors, however, seem to be spatially independent. The spatial correlation coefficients of the errors (calculated based on the construction of the correlogram of the estimation errors) are only at the order of 0.01.


Journal of Geographic Information and Decision Analysis, Vol. 2., No. 2, pp. 243-251, 1998.

Keywords: Shannon's Entropy, Maximum Entropy Principle, Lognormal Random Field, Simple Kriging, Log-Kriging.
Topic revision: r3 - 13 Aug 2010 20:27:13, theresiafreska
This site is powered by FoswikiCopyright &© by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Wiki? Send feedback