Re: correlation coefficient

Roger Stafford wrote:


In article <ef4033e.4@xxxxxxxxxxxxxxxxxxxxxxx>, Eskofier
<bjoern@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:

When created with randn, corralation should indeed be zero.

Read the documentation for corr, too. It computes the Spearman,
Kendall, and by default the Pearson correlation.
-----------------------
Ideally, that would be true, Eskofier, but what we're dealing
with here
is a finite sample, so the sample-derived correlation will
generally
differ from zero. This is no fault of matlab's random functions,
but a
basic characteristic in statistics. The correlation between two
independent random variables derived from a finite sample will not
be zero
in general, but has a well-defined non zero variance which depends
on the
size of the sample.

To illustrate that, when I did the following four times,

c = corrcoef(randn(4096,1),randn(4096,1)),

the four answers for c(1,2) I got were 0.0427, 0.0023, 0.0195, and
0.0122,
which are small but not quite zero. If we used Nivas' 10000, these
would
be smaller but still not zero.

Roger Stafford


Go on and calculate the ML Estimator. You'd find it to be 0.
.

Relevant Pages

  • Re: correlation coefficient
    ... In article, Eskofier ... and by default the Pearson correlation. ... differ from zero. ... independent random variables derived from a finite sample will not be zero ...
    (comp.soft-sys.matlab)
  • Re: Correlation of X with XY ?
    ... For Mean=0 the correlation will in fact be zero. ... vectors fall on the "13 symmetry axes of a cube ". ... a dispute arose because actual Psychology Tests ...
    (sci.stat.math)
  • Re: polychoric correlations
    ... As Rich said, with binary items, the term "tetrachoric correlation" is ... If a single cell of the fourfold table has a zero frequency, ...
    (sci.stat.consult)
  • Re: How to create conditional correlated random values given exogenous data
    ... Just exactly what IS a correlation structure? ... If I tell you that X and Y have ZERO correlation, ... The scatter of residuals versus fitted is anything BUT points ... -- Reef Fish Bob. ...
    (sci.stat.math)
  • Re: Rounding problem with corrcoef() ?
    ... It seems that Matlab is not interpreting an ... operation that results in an answer of zero to be zero. ... Since all the values of x1 are the same, the correlation ...
    (comp.soft-sys.matlab)