Information Technology Reference
In-Depth Information
6 apply crossover to generate an off spring (
child)
7 mutate
child
8
If
dist
(
child, parent
1)
<
>
dist
(
child, parent
2)
^
fi tness(
child
)
fi tness(
parent
1)
←
10
h en
parent
1
child
>=
11
ElseIf
dist
(
child, parent
1)
dist
(
child, parent
2)
>
12
^
fi t n e s s
(
child
)
fi t n e s s
(
parent
2)
13
h en
parent
2
←
child
14
EndIf
15
EndFor
16
EndFor
17 extract the best individuals from the population and add them
to the fi nal solution
18
EndFor
4.6.2
Negative Selection with Fuzzy Detection Rules
Gonzalez (2003) applied fuzzy rules instead of crisp rules for detectors. h at is,
given a set of self-samples, the algorithm will generate fuzzy detection rules in the
nonself space that can determine if a new sample is normal or abnormal. Results
have shown that the use of fuzzy rules improves the accuracy of the method and
produces a measure of deviation from the normal that does not need to partition
the nonself space.
A fuzzy detection rule has the following structure:
If
xT
∈∧
…
x T
n
∈
then
non_self
1
1
n
where
=
(x
1
,
…
,
x
n
)
element of the self/nonself space being evaluated
=
T
i
fuzzy set
^
=
fuzzy conjunction operator (in this case,
min
)
h e fuzzy set
T
i
is defi ned by a combination of basic fuzzy sets (linguistic values).
Given a set of linguistic values
S
i
⊆
=
{
S
1
, …,
S
m
}
and subset
TS
associated to
each fuzzy set
T
i
,
∈
∪
T
S
i
j
ST
j
i
∪
where
corresponds to a fuzzy disjunction operator (here addition operator)
defi ned as follows:
()
x
min{ ()
x
(),}
x
1
∪
AB
A
B
Search WWH ::
Custom Search