Abstract: This paper researches the problem of restoring the functions-signals which allow a spectral representation by a finite set of data which is corrupted by errors. The authors consider the cases when the errors are random values and which certain statistical information is either known or not (the errors are considered as fuzzy values). The pseudoinverse solution, solution in the form of an expansion by eigenfunctions as well as Pareto-optimal and the maximum plausible estimations are constructed. A connection between them is established.

Keywords: function-signal, pseudoinversion, maximum plausible estimations, Pareto-optimization, fuzzy values.

ACM Classification Keywords: I.6 Simulation and Modeling.

Link:

PROBLEM OF RESTORING THE FUNCTIONS-SIGNALS BY FINITE SET OF DATA WITH ERRORS

Veda Kasyanyuk, Iryna Volchyna

http://www.foibg.com/ijima/vol02/ijima02-04-p07.pdf