Search

Results:

Vol. 7 (2004 year), No. 3

Panteleev V.P.
The best L₁-approximation in classes of distributions

The paper has considered a gradient method of finding the best L₁-approximation in classes of distributions. The method is based on geometrical differentiation and can be applied for approximation of functions, expressed in a graph form, for instance, by reading out of devices. The method, expounded here for the class of normal distributions, can readily be extended to other classes of distributions.

(in Russian, стр.6, fig. 2, tables. 0, ref 1, MS Word 95, MS Word 95 41 Kb)