ccf_errors: Use random simulations to estimate undertainty on CCF...

In svdataman/sour: Cross-correlation of time series (which may be unevenly sampled)

Description

ccf_errors returns information on the undertainty on CCF estimates.

Usage

ccf_errors(ts.1, ts.2, tau = NULL, min.pts = 5, local.est = FALSE,
  cov = FALSE, prob = 0.1, nsim = 250, peak.frac = 0.8,
  zero.clip = NULL, one.way = FALSE, method = "iccf",
  use.errors = FALSE, acf.flag = FALSE, chatter = 0)

Arguments

ts.1	(array or dataframe) data for time series 1 and 2.
ts.2	(array or dataframe) data for time series 1 and 2.
tau	(array) list of lags at which the CCF is to be evaluated.
min.pts	(integer) each DCF bin must contain at least min.pts correlation coefficients.
local.est	(logical) use 'local' (not 'global') means and variances?
cov	(logical) if TRUE then compute covariance, not correlation coefficient.
prob	(logical) probability level to use for confidence intervals
nsim	(integer) number of FR/RSS simulations to run
peak.frac	(float) only include CCF points above peak.frac*max(ccf) in centroid calculation.
zero.clip	(logical) remove pairs of points with exactly zero lag?
one.way	(logical) (ICCF only) if TRUE then only interpolar time series 2.
method	(string) use "dcf" or "iccf" (default).
use.errors	(logical) if TRUE then subtract mean square error from variances.
acf.flag	(logical) TRUE when computing ACF, and ts.2 = ts.1
chatter	(integer) set the level of feedback.

Details

Computes errors on the CCF estimates using "flux randomisation" and "random subset sampling" FR/RSS using the fr_rss function.

Value

The output is a list containing two data frames: lags and dists.

lags	a data frame with four columns
tau	time lags
dcf	the DCF values for the input data
lower	the lower limit of the confidence interval
upper	the upper limit of the confidence interval
dists	a data frame with two columns
peak.lag	the peak values from nsim simulations
cent.lag	the centroid values from nsim simulations

Notes

For each randomised pair of light curves we compute the CCF. We record the CCF, the lag at the peak, and the centroid lag (including only points higher than peak.frac * max(ccf)). Using nsim simulations we compute the (1-p)*100% confidence intervals on the CCF values, and the distribution of the peaks and centroids.

See Also

cross_correlate, fr_rss

svdataman/sour documentation built on May 30, 2019, 8:47 p.m.