Math Forum Discussions

Derek Goring

Posts: 3,644
Registered: 12/7/04

Re: Correlation Coefficients
Posted: Mar 15, 2013 5:20 PM

Plain Text

On Saturday, March 16, 2013 2:26:07 AM UTC+13, Lola wrote:
> Hello,
>
> I've obatin the correlation coefficient of two time series using xcorr. The best correlation coefficient has a time lag of 2 month. I want to estimate the significance level and the confidence interval. The only way I found to get the significance level is by the p-value of the corrcoef but it needs time series without time lag. Does anyone have an idea?
>
> Thanks

So, you need to time-lag one of the vectors before using corrcoef.
[R,p]=corrcoef(x(1:end-lag),y(lag+1:end));