We discuss pairwise algorithms, a kind of computational algorithm which can be useful in dynamically updating statistics as new samples of data are collected. Since test data are usually collected through time as individual data sets, these algorithms would be profitably used in computer programs to treat this situation. Pair-wise algorithms are presented for calculating the sum of products of deviations about the mean for adding a sample of data (or removing one) to the whole data set.
Get full access to this article
View all access options for this article.
References
1.
ChanT. F.GolubG. H.LevequeR. J. (1983) Algorithms for computing the sample variance: analysis and recommendations. The American Statistician, 37, 242–247.
2.
CottonE. W. (1975) Remark on Stably updating mean and standard deviation of data. Communications of the ACM, 18, 458.
3.
GrantI. H. W. M. (1987) Recursive least square. Teaching Statistics, 9, 15–18.
4.
SearleS. R. (1983) The recurrence formulae for means and variances. Teaching Statistics, 5, 7–10.
5.
Van ReekenA. J. (1968) Dealing with Neely's algorithms. Communications of the ACM, 3, 149–150.
6.
WelfordB. P. (1962) Note on a method for calculating corrected sums of squares and products. Technometrics, 4, 419–420.
7.
YoungsE. A.CramerE. M. (1971) Some results relevant to choice of sum and sum-of-product algorithms. Technometrics, 13, 657–665.
