IEEE Transactions on Fuzzy Systems, Год журнала: 2024, Номер 32(8), С. 4285 - 4296
Опубликована: Апрель 25, 2024
Anomaly
detection
is
a
significant
area
of
discovering
knowledge
that
has
shown
success
in
the
areas
fraud
detection,
cyber
security,
and
medical
diagnostics.
The
kernelized
fuzzy-rough
set
key
extension
model
rough
computing.
It
inherits
advantages
kernel
function
can
handle
uncertain
information
data
more
effectively.
However,
existing
models
construct
upper
lower
approximation
sets
mainly
from
decision
attribute
which
are
not
available
to
unlabeled
data.
In
addition,
best
our
knowledge,
studies
related
use
fuzzy
for
constructing
effective
anomaly
have
yet
been
reported.
Based
on
these
observations,
this
paper
constructs
proposes
(KFRAD)
method.
Specifically,
we
first
optimize
compute
relation
matrix
subsets
space.
Then,
definition
accuracy
given.
granule
extent
determined
based
accuracy.
Finally,
extents
corresponding
weight
values
integrated
scores
objects.
On
basis
above
ideas,
design
KFRAD
algorithm
experimentally
compare
it
with
mainstream
algorithms.
analysis
results
show
proposed
better
performance.
code
publicly
online
at
Язык: Английский