An inductive algorithm is presented for smooth approximation of functions based on the Tikhonov regularization method. The discrepancy principle is used for estimating the optimal value of the smoothing parameter. The principle of heuristic self-organization is applied for assessment of some parameters and the optimal complexity of the approximating function. In this paper the efficiency of this algorithm was investigated for different kind of the Tikhonov parametric functional using model functions distorted by additive Gaussian noise simulated by a random numbers generator.
REFERENCES
1.
A.
Talmi
and G.
Gilat
., J. Comp. Phys
23
, 93
–123
(1977
).2.
A.G.
Ivakhnenko
, Automatica
, 6
, 207
–210
(1970
).3.
Beer
S.
, Cybernetics and Management
(The English Universities Press LTD
, London
) (1967
).4.
V.
Stoyanova
, D.
Kashchiev
and T.
Kupenova
, J. Aerosol Sci.
, 25
, No. 5
, 867
–877
(1994
).5.
T.
Kupenova
and P.
Kupenova
in Proceedings of 5-th General Conference of the Balkan Physical Union BPU5
, edited by. S. Jokicet et al., (Serbian Physical Society
, Belgrade
), 1475
–1478
(2003
).6.
V.
Stoyanova
and T.
Kupenova
, Compt. Rend. Acad. Bulg. Sci.
, 60
, No. 5
, 545
–550
(2007
).7.
P.
Kupenova
, E.
Popova
, and L.
Vitanova
. Vision Res.
, 48
, 882
–892
(2008
).8.
9.
This content is only available via PDF.
© 2019 Author(s).
2019
Author(s)
You do not currently have access to this content.
