cor_ts {BINCOR}R Documentation

Bi-variate correlation

Description

The cor_ts function estimates the correlation between the binned time series. cor_ts estimates three types of correlation coefficients: Pearson’s correlation, Spearman’s and Kendall’s rank correlations by means of the R native function cor.test (package:stats). The cor_ts function has an option to remove the linear trend of the time series under analysis (other pre-processing methods could be used) and its parameters are described in the following lines.

Usage

cor_ts(bints1, bints2, varnamets1="NULL", varnamets2="NULL", KoCM, rmltrd="N", 
 device="screen", Hfig, Wfig, Hpdf, Wpdf, resfig, ofilename)

Arguments

bints1, bints2

The bints1 and bints2 are the binned time series.

varnamets1, varnamets2

varnamets[1][2] are the names of the variables under study.

KoCM

KoCM indicates the correlation estimator: pearson for Pearson (the option by default), spearman for Spearman and kendall for Kendall.

rmltrd

This is the option used to remove the linear trend in the time series under study (by default the linear trend is not removed, but it can be activated with the option “Y” or “y”).

device

The type of the output device (by default the option is “screen”, and the other options are “jpg”, “png” and “pdf”) for the scatter plot for the binned time series.

Hfig

The height for the scatter plot in “jpg” or “png” format.

Wfig

The width for the scatter plot in “jpg” or “png” format.

Hpdf

The height for the scatter plot in “pdf” format.

Wpdf

The width for the scatter plot in “pdf” format.

resfig

resfig is the resolution in “ppi” (by default R does not record a resolution in the image file, except for BMP) for the scatter plot (“jpg” or “png” formats), an adequate value could be 150 ppi.

ofilename

The output filename for the scatter plot of the binned time series.

Details

The cor_ts estimate the correlation between two binned time series by means of the R native function cor.test (package:stats).

Value

Output: an object of the form cor.test containing the correlation coefficient and the statistical significance.

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

Author(s)

Josué M. Polanco-Martínez (a.k.a. jomopo).
BC3 - Basque Centre for Climate Change, Bilbao, SPAIN.
EPOC UMR CNRS 5805 - U. de Bordeaux, Pessac, FRANCE.
Web1: https://scholar.google.es/citations?user=8djLIhcAAAAJ&hl=en.
Web2: http://www.researchgate.net/profile/Josue_Polanco-Martinez.
Email: josue.m.polanco@gmail.com

References

Mudelsee, M. (2010). Climate Time Series Analysis: Classical Statistical and Bootstrap Methods. Springer.

Mudelsee, M. (2014). Climate Time Series Analysis: Classical Statistical and Bootstrap Methods, Second Edition. Springer.

Polanco-Martínez, J.M., Medina-Elizalde, M.A., Sánchez Goñi, M.F., M. Mudelsee. (2018). BINCOR: an R package to estimate the correlation between two unevenly spaced time series. Ms. under review (second round).

Examples

 #####################################################################
 #::  Figure 2 (Polanco-Martínez et al. (2018), (mimeo)). 
 #####################################################################
 library("BINCOR")  
 library("pracma")

 #####################################################################
 #:: Loading the time series under analysis: example 1 (ENSO vs. NHSST) 
 #####################################################################
 data(ENSO) 
 data(NHSST)

 #####################################################################
 # Computing the binned time series though our bin_cor function 
 #####################################################################
 bincor.tmp    <- bin_cor(ENSO.dat, NHSST.dat, FLAGTAU=3, "output_ENSO_NHSST.tmp")
 binnedts      <- bincor.tmp$Binned_time_series

 #####################################################################
 # Testing our cor_ts function: cor_ts.R
 #####################################################################
 # screen (scatterplot) and Pearson  
 cor_ts(binnedts[,1:2], binnedts[,c(1,3)], "ENSO-Nino3", "SST NH Mean", 
  KoCM="pearson", rmltrd="y") 

 # PDF format (scatterplot) and Kendall 
 cor_ts(binnedts[,1:2], binnedts[,c(1,3)], "ENSO-Nino3", "SST NH Mean", 
  KoCM="kendall", rmltrd="y", device="pdf", Hpdf=6, Wpdf=9, resfig=300, 
  ofilename="scatterplot_ENSO_SST") 

 # JPG format (scatterplot) and Spearman  
 cor_ts( binnedts[,1:2], binnedts[,c(1,3)], "ENSO-Nino3", "SST NH Mean", 
  KoCM="spearman", rmltrd="y", device="jpg", Hfig=900, Wfig=1200, 
  resfig=150, ofilename="scatterplot_ENSO_SST") 

[Package BINCOR version 0.2.0 Index]