dcf

       compute the Edelson & Krolik discrete correlation function

       time series analysis tools

       dcf,ta,ra,tb,rb,lag,cor,numf=numf, $
           erra=erra_in,errb=errb_in, $
           numpt=numpt,minpt=minpt, $
           minlag=minlag,maxlag=maxlag

       ta,ra: time and rate of the first light curve
       tb,rb: time and rate of the second light curve

       erra, errb: Arrays containing the measurement errors of
             ra and rb, respectively. If given, this is taken into
             account by using eq. (3) of Edelson & Krolik.
             NOTE: There are lots of problems with interpreting the
             DCF computed this way instead of using Edelson & Krolik,
             eq. (2), and it is NOT recommended to give erra and errb
             (see, e.g., White & Peterson, 1994, PASP, 106, 879)
       minlag: minimum lag to consider
       maxlag: maximum lag to consider
       numf:   number of lag bins for which to compute the DCF
       minpt:  minimum number of data points for a DCF value to be
               computed (default: 10)

       lag: lag value
       cor: value of the DCF, or NAN if it cannot be determined
            (<minpt )

       numpt:  number of data points entering each lag estimate

     The lightcurves should be stationary, i.e., prewhitening by
     subtracting off a low order polynomial from the data might be
     required. See Welsh (1999, PASP 111, 1347)

     Implementation of the formulae of Edelson & Krolik, ApJ 333,
     646, 1988. You should read this paper as well as the papers
     cited above before using this procedure.

 $Log: dcf.pro,v $
 Revision 1.3  2005/11/17 14:17:59  wilms
 force computation in double for means and standard deviations to make
 computations more robust.
 Revision 1.2  2005/11/17 12:36:47  wilms
 bug fix: the first cor bin could include data not within the
 respective lag bin
 Revision 1.1  2005/11/17 12:24:58  wilms
 initial release

[Home Page] [Software, Documentation] [IDL Documentation] [Quick Reference] [Feedback]