Information Technology Reference
In-Depth Information
Fig. 10 Input Output structure of the PV production forecast neural network
3. Apply the criterion for adding a new hidden unit:
if
e(n)
> E 1 and e rms (n)>E 2 and d(n)>E 3 allocate a new hidden unit K +1
with:
￿ ʻ K+1 = e(n)
￿
c K þ 1 ¼
x
ð
n
Þ
￿ b K þ 1 ¼ a
k
x
ð
n
Þ
c r ð
n
Þ
k
else
tune the network parameters:
￿
w
ð n Þ ¼
w
ð n
1
Þþ
k
ð n Þ e ð n Þ
update the error covariance matrix:
￿
P
T
P
ð
n
Þ ¼
I
k
ð
n
Þ a
ð
n
Þ
ð
n
1
Þþ
Q
ð
n
1
Þ
end
4. Check the criterion to prune hidden units:
compute the hidden unit outputs:
￿
o i ð
n
Þ ¼k i /
ð
k
x
ð
n
Þ
c i
k
Þ;
i
¼
1
; ...;
K
compute the normalized outputs:
￿
o i ð
n
Þ
o i ð
n
Þ ¼
Þg ;
i
¼
1
; ...;
K
f
o i ð
max
n
Search WWH ::




Custom Search