R Graphical Manual

Browse All

Last data update: 2014.03.03

R: Scale decomposition for polygon data.

iBEEMD

R Documentation

Scale decomposition for polygon data.

Description

Similar to 2-dimensional wavelet decomposition, for a given irrelular-lattice field represented by spatial polygons dataframe, the method decompose the field into different scales and a global trend component by EEMD method. The scale components are also called also called intrinsic mode functions (IMFs), which represent different scale information in the spatial field.

Usage

iBEEMD(spPolysDf, valueField = names(spPolysDf)[1], nMaxIMF = 10, tolSift = 0.05, 
neemd = 1000, wnsd = 0.05, fmodel = "thinplate", fig = TRUE)

Arguments

`spPolysDf`	a SpatialPolygonsDataFrame object.
`valueField`	a field name that stores value.
`nMaxIMF`	maximum number of components to be decomposed.
`tolSift`	sift tolerence, a small number.
`neemd`	number of EEMD iterations, a large number can make a stable result.
`wnsd`	standard deviation of added noise; it is a ratio to the standard deviation of above data.
`fmodel`	surface fitting function ("thinplate", "gaussian", "cubic", "multiquadric").
`fig`	whether plot decomposed results.

Value

A SpatialPolygonsDataFrame with original value, decomposed imfs and global trend.

References

Hu, M.-G. and J.-F. Wang, et al. A hierarchical-scale decomposition method for irregular lattice field. Computers & Geosciences, submitted.

Huang, N. E. and Z. Shen, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis. Proceedings of The Royal Society A - Mathematical, Physical & Engineering Sciences, 1998, 454(1971): 903-995.

Examples

## Not run: 
library(rgdal)

# polygon data
mydata <- system.file("extdata/simu.shp", package = "ibeemd")

layer <- basename(mydata)
layer <- substr(layer, 1, nchar(layer)-4)
mydataDf <- readOGR(dsn=mydata, layer=layer)
#spplot(mydataDf)

rslt <- iBEEMD(
		spPolysDf = mydataDf, 
		valueField = "value",  
		nMaxIMF = 10, 
		tolSift = 0.05,
		neemd = 500,
		wnsd = 0.05,
		fmodel = "thinplate",
		fig = TRUE)

## End(Not run)
#spplot(rslt)