UnivariateStandardDeviation Method |
Namespace: Meta.Numerics.Statistics
Exception | Condition |
---|---|
ArgumentNullException | sample is . |
Note that a few authors use the term "sample standard deviation" to mean "the square root of the population variance as estimated by the sample". We do not adopt this aberrant convention, which is contrary to the conventional meaning of "sample moment" for all other moments. In other words, what is returned by this methods is the square root of the sum of the squared deviations divided by n, not divided by (n-1). If you want the square root of the sum of the squared deviations divided by n, which is also called the Bessel-corrected standard deviation, use the CorrectedStandardDeviation(IReadOnlyCollectionDouble) method. Note, however, that, unlike the variance, the standard deviation corrected in this way is not unbiased. If you want a bias-corrected estimate of the standard deviation of the populace that was sampled, use the PopulationStandardDeviation(IReadOnlyCollectionDouble) method.