Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: Significance Levels for Correlation Functions
Posted:
Jan 20, 2013 9:00 AM


Greg Heath <g.heath@verizon.net> wrote in message <aa1a73b0e4754c56a37843bd1266ef63@y5g2000pbi.googlegroups.com>... > I standardize (zeromean/unitbiasedvariance) before calculating > autocorr(x) and > crosscorr(x,y). The results are biased estimates with unity at zero > lag for the autocorrelation. > > To obtain unbiased estimates divide by N1 instead of N for the > variance and > divide by Nabs(k)1 instead of N for the kth lag. I didn't like the > resulting plots, > so I use the biased estimate. > > To determine significance levels I averaged over M=100 trials for N = > 100 dimensional Normal Gaussian time series. The 95th average absolute > value was 0.21 and the 100th was 3.1. Therefore I consider correlation > values >= 0.21 as significant. > > A noisefree nth order polynomial can be determined by n+1 points. > Therefore > when I imagine a nth order polynomial fit to a smoothed plot of x, or > y vs x, I start thinking about nonzero lags <= n+1. > > Hope this helps. > > Greg
sigthresh95(N) is a function of N.
I repeated the N=100 experiment for M = 100 and M=1000. I now get
sigthresh95(100) ~ 0.15 %(NOT 0.21 !)
Hope this helps as a check for your calculations.
Greg



