Chronology | Current Month | Current Thread | Current Date |
[Year List] [Month List (current year)] | [Date Index] [Thread Index] | [Thread Prev] [Thread Next] | [Date Prev] [Date Next] |
On 10/09/2006 05:25 PM, Krishna Chowdary wrote:
Are we talking about the sample variance versus population variance?
If so, the fix is simple: replace N by N-1 in the normalization.
That allows you to convert the sample variance into an unbiased predictor
of the population variance.
http://mathworld.wolfram.com/SampleVariance.html
But perhaps I have misunderstood the question. The generally-used
notation
is so messed up that all sorts of inconsistencies can arise.