TITLE:
Application of Multi-Gene Genetic Programming in Kriging Interpolation
AUTHORS:
Changik Han, Ende Wang, Jianming Xia, Sunggi Yun
KEYWORDS:
MGGP, Kriging Interpolation, Variogram
JOURNAL NAME:
Journal of Geoscience and Environment Protection,
Vol.3 No.5,
July
17,
2015
ABSTRACT: A key stage for Kriging interpolation is
the estimating of variogram model, which characterizes the spatial behavior of
the variables of interest. But most traditional kriging interpolation has
finite types of empirical variogram model, and sometimes, the optimal type of
variogram model can not be find, which result in decreasing interpolation
accuracy. In this paper, we explore the use of Multi-Gene Genetic Programming
(MGGP) to automatically find an empirical variogram model that fits on an
experimental variogram. Empirical variogram estimation based on MGGP, in
contrast with traditional method need not select type of basic variogram model
and can directly get both the functional type as well as the coefficients of
the optimal variogram. The results of case study show that the proposed method
can avoid the subjectivity in choosing the type of variogram models and can
adaptively fit variogram according to the real data structure, which improves
the interpolation accuracy of kriging significantly.