getSL: Estimate correlation significance levels using...

In hotblack43/PTHfftSURROGATES: Establishes correlation significance levels using phase scrambling methods.

Description

Use the phase_scramble function to generate many (Nboot) surrogate-data based correlations, compare to the observed correlation, calculate the fraction of trials exceeding the observed value and return the significance level as a fraction.

Usage

getSL(Nboot,method,series1,series2)

Arguments

Nboot	Number of Monte-Carlo phase-scrambling trials to perform
method	pearson, kendall or spearman
series1,series2	the two series of numbers to correlate

Value

SL	Significance levels, or fraction of surrogat edata trials exceeding the observed correlation. Two sided!

Note

Version 0.1, Dec 2019

Author(s)

Peter Thejll, DMI. ([e-mail])

Examples

library(PTHfftSURROGATES)
require(gtools)
set.seed(6) # for reproducibility

#Example 1 - two probably uncorrelated series
series1 <- rnorm(1000)
series2 <- rnorm(1000)
method_str="pearson" # pearson, kendall or spearman
Robs <- cor(series1,series2,method=method_str)
print(c("The observed correlation is ",Robs))
Nboot <- 1000
SL <- getSL(Nboot,method=method_str,series1,series2)
print(c("The p-value of the observed correlation is:",SL))

#Example 2 - two probably correlated series
series2 <- 2*rnorm(1000) + 1*series1
Robs <- cor(series1,series2,method=method_str)
print(c("The observed correlation is ",Robs))
SL <- getSL(Nboot,method=method_str,series1,series2)
print(c("The p-value of the observed correlation is",SL))

hotblack43/PTHfftSURROGATES documentation built on Dec. 25, 2019, 6:15 p.m.