Neurocomputing in Complex Domain - High Dimensional Neurocomputing: Growth, Appraisal and Applications - page 63

Fig. 3.10 a Plane and b Surface graphs for Mean-Median Error Function

n (

E = sinh ( e i )) (3.15)

The complex sine-hyperbolic EF is defined to be

=

( | ʵ i | )

E sinh (3.16)

n

where n is the number of outputs (Fig. 3.11 ).

3.3.2.7 Cauchy Error Function

Cauchy EF, also known as the Lorentzian function, is one of the most robust functions
of statistics. The tuning constant (c) for the EF is 2.3849. The cauchy's EF is given
by

ln 1

2

e i
c

c 2
2

E = + (3.17)

n

The complex Cauchy EF is defined to be

ln 1

ʵ i ʵ i
c 2

c 2
2

E = + (3.18)

n