WC {W2CWM2C}

Wavelet correlation (bivariate case) pairwise comparisons.

Description

The WC function (bivariate case) computes the wavelet correlation by means of the function wave.-correlation of the waveslim package to several time series, makes a pairwise comparisons and plot the pairwise wavelet correlations in descending order as a single heatmap using the colorspace package. The input data are multivariate time series and WC function only tackle arrays with N x C (elements x columns, where the number of columns are between 2 and 7) dimensions.

Usage

WC(inputDATA, Wname, J, device="screen", filename, 
   Hfig, WFig, Hpdf, Wpdf)

Arguments

inputDATA

An array of multivariate time series as a ts object (please, check the ts manual to get more information about the ts function in R).

Wname

The wavelet function or filter to use in the decomposition.

J

Specifies the depth of the decomposition.

device

The type of the output device (by default the option is “screen”, and the other options are “jpg”, “png”, “eps” and “pdf”).

filename

The output filename.

Hfig

The height of the 'jpg' or 'png' image.

WFig

The width of the 'jpg' or 'png' image.

Hpdf

The height of the 'eps' or 'pdf'.

Wpdf

The width of the 'eps' or 'pdf'.

Details

The WC function compute the wavelet correlation among time series and plots the results in a single heatmap plot (which can be displayed in the screen or can be saved as PNG, JPG, PDF or EPS) showing the WC values as a table (please, look at Figure 3 in Polanco-Martinez and Fernandez-Macho 2014). The WC code is based on the wave.correlation routine from Brandon Whitcher's waveslim R package Version: 1.7.1, which is based mainly on wavelet methodology developed by Whitcher, B., P. Guttorp and D.B. Percival (2000) and Gencay, Selcuk and Whitcher (2001).

Value

Output:

Output plot: screen or 'filename + .png, .jpg, .eps or .pdf'.

wavcor.modwtsDAT: matrix with as many rows as levels in the wavelet transform object. The first column provides the point estimate for the wavelet correlation followed by the lower and upper bounds from the confidence interval.

to3DpL: A matrix (the matrix table added in the WC plot) with a J (number of wavelet scales) X C (the number of pairwise comparisons) dimensions, which are in descending order taking into account the sum of the wavelet correlation coefficients for all (J) wavelet scales.

Note

Needs waveslim package to calculate modwt, brick.wall and the wave.correlation and also needs the colorspace package to plot the heatmaps.

Author(s)

Josue M. Polanco-Martinez (a.k.a. jomopo).
BC3 - Basque Centre for Climate Change, Bilbao, Spain.
Web1: https://scholar.google.es/citations?user=8djLIhcAAAAJ&hl=en.
Web2: https://www.researchgate.net/profile/Josue_Polanco-Martinez.
Email: [e-mail].

References

Gencay, R., F. Selcuk and B. Whitcher (2001). An Introduction to Wavelets and Other Filtering Methods in Finance and Economics, Academic Press.

Ihaka, R., Murrell, P., Hornik, K., Fisher, J. C. and Zeileis, A. (2012). colorspace: Color Space Manipulation. R package version 1.2.0, The Comprehensive R Archive Network (CRAN), <URL: https://cran.r-project.org/package=colorspace>.

Polanco-Martinez, J. and J. Fernandez-Macho (2014). The package 'W2CWM2C': description, features and applications. Computing in Science & Engineering, 16(6):68–78.
doi: 10.1109/MCSE.2014.96.

Whitcher, B., P. Guttorp, and D.B. Percival (2000). Wavelet analysis of covariance with application to atmospheric time series. Journal of Geophysical Research - Atmospheres, 105(D11):941–962. doi: 10.1029/2000JD900110.

Whitcher, B. (2012). waveslim: Basic wavelet routines for one-, two- and three-dimensional signal processing. R package version 1.7.1, The Comprehensive R Archive Network (CRAN),
<URL https://cran.r-project.org/package=waveslim>.

Examples

## Figure 3 (Polanco-Martinez and Fernandez-Macho 2014). 

 library("colorspace")
 library("waveslim")
 library("W2CWM2C") 
 data(dataexample)   

 #:: Transforms to log returns using: ln(t + deltat) - ln(t). 
 #:: The application in this example uses stock market 
 #:: indexes (it is common to use log returns instead of
 #:: raw data). Other kinds of pre-processing data are possible. 

 dataexample  <- dataexample[-1] # remove dates!
 dataexample  <- dataexample[,1:5]
 lrdatex      <- apply(log(dataexample), 2, diff)
 inputDATA    <- ts(lrdatex, start=1, frequency=1) 

 #Input parameters 
  Wname       <- "la8"
  J           <- 8 
  compWC      <- WC(inputDATA, Wname, J, device="screen", NULL,
                  NULL, NULL, NULL, NULL)