<s>
Some	O
of	O
these	O
methods	O
include	O
knowledge	B-Application
discovery	I-Application
in	I-Application
databases	I-Application
(	O
KDD	O
)	O
,	O
data	B-Application
mining	I-Application
,	O
machine	O
learning	O
and	O
statistics	O
.	O
</s>
<s>
In	O
order	O
to	O
effectively	O
test	O
,	O
detect	O
,	O
validate	O
,	O
correct	O
error	O
and	O
monitor	O
control	O
systems	O
against	O
fraudulent	O
activities	O
,	O
businesses	O
entities	O
and	O
organizations	O
rely	O
on	O
specialized	O
data	O
analytics	O
techniques	O
such	O
as	O
data	B-Application
mining	I-Application
,	O
data	O
matching	O
,	O
the	O
sounds	O
like	O
function	O
,	O
regression	O
analysis	O
,	O
clustering	B-Algorithm
analysis	O
,	O
and	O
gap	O
analysis	O
.	O
</s>
<s>
Techniques	O
used	O
for	O
fraud	O
detection	O
fall	O
into	O
two	O
primary	O
classes	O
:	O
statistical	O
techniques	O
and	O
artificial	B-Application
intelligence	I-Application
.	I-Application
</s>
<s>
Data	B-General_Concept
preprocessing	I-General_Concept
techniques	O
for	O
detection	O
,	O
validation	O
,	O
error	B-Error_Name
correction	I-Error_Name
,	O
and	O
filling	O
up	O
of	O
missing	O
or	O
incorrect	O
data	O
.	O
</s>
<s>
Clustering	B-Algorithm
and	O
classification	B-General_Concept
to	O
find	O
patterns	O
and	O
associations	B-Algorithm
among	O
groups	O
of	O
data	O
.	O
</s>
<s>
Matching	O
algorithms	O
to	O
detect	B-Algorithm
anomalies	I-Algorithm
in	O
the	O
behavior	O
of	O
transactions	O
or	O
users	O
as	O
compared	O
to	O
previously	O
known	O
models	O
and	O
profiles	O
.	O
</s>
<s>
Techniques	O
are	O
also	O
needed	O
to	O
eliminate	O
false	B-Error_Name
alarms	I-Error_Name
,	O
estimate	O
risks	O
,	O
and	O
predict	O
future	O
of	O
current	O
transactions	O
or	O
users	O
.	O
</s>
<s>
Some	O
forensic	O
accountants	O
specialize	O
in	O
forensic	O
analytics	O
which	O
is	O
the	O
procurement	O
and	O
analysis	O
of	O
electronic	B-General_Concept
data	I-General_Concept
to	O
reconstruct	O
,	O
detect	O
,	O
or	O
otherwise	O
support	O
a	O
claim	O
of	O
financial	O
fraud	O
.	O
</s>
<s>
The	O
main	O
steps	O
in	O
forensic	O
analytics	O
are	O
data	O
collection	O
,	O
data	B-General_Concept
preparation	I-General_Concept
,	O
data	O
analysis	O
,	O
and	O
reporting	O
.	O
</s>
<s>
The	O
main	O
AI	B-Algorithm
techniques	I-Algorithm
used	I-Algorithm
for	I-Algorithm
fraud	I-Algorithm
detection	I-Algorithm
include	O
:	O
</s>
<s>
Data	B-Application
mining	I-Application
to	O
classify	O
,	O
cluster	O
,	O
and	O
segment	O
the	O
data	O
and	O
automatically	O
find	O
associations	B-Algorithm
and	O
rules	O
in	O
the	O
data	O
that	O
may	O
signify	O
interesting	O
patterns	O
,	O
including	O
those	O
related	O
to	O
fraud	O
.	O
</s>
<s>
Expert	B-Application
systems	I-Application
to	O
encode	O
expertise	O
for	O
detecting	O
fraud	O
in	O
the	O
form	O
of	O
rules	O
.	O
</s>
<s>
Neural	B-Architecture
nets	I-Architecture
to	O
independently	O
generate	O
classification	B-General_Concept
,	O
clustering	B-Algorithm
,	O
generalization	O
,	O
and	O
forecasting	O
that	O
can	O
then	O
be	O
compared	O
against	O
conclusions	O
raised	O
in	O
internal	O
audits	O
or	O
formal	O
financial	O
documents	O
such	O
as	O
10-Q	O
.	O
</s>
<s>
Other	O
techniques	O
such	O
as	O
link	B-Application
analysis	I-Application
,	O
Bayesian	O
networks	O
,	O
decision	O
theory	O
,	O
and	O
sequence	B-Language
matching	I-Language
are	O
also	O
used	O
for	O
fraud	O
detection	O
.	O
</s>
<s>
If	O
data	B-Application
mining	I-Application
results	O
in	O
discovering	O
meaningful	O
patterns	O
,	O
data	O
turns	O
into	O
information	O
.	O
</s>
<s>
The	O
machine	O
learning	O
and	O
artificial	B-Application
intelligence	I-Application
solutions	O
may	O
be	O
classified	O
into	O
two	O
categories	O
:	O
'	O
supervised	O
 '	O
and	O
'	O
unsupervised	O
 '	O
learning	O
.	O
</s>
<s>
Supervised	O
neural	B-Architecture
networks	I-Architecture
,	O
fuzzy	O
neural	B-Architecture
nets	I-Architecture
,	O
and	O
combinations	O
of	O
neural	B-Architecture
nets	I-Architecture
and	O
rules	O
,	O
have	O
been	O
extensively	O
explored	O
and	O
used	O
for	O
detecting	O
fraud	O
in	O
mobile	O
phone	O
networks	O
and	O
financial	O
statement	O
fraud	O
.	O
</s>
<s>
Bayesian	O
learning	O
neural	B-Architecture
network	I-Architecture
is	O
implemented	O
for	O
credit	O
card	O
fraud	O
detection	O
,	O
telecommunications	O
fraud	O
,	O
auto	O
claim	O
fraud	O
detection	O
,	O
and	O
medical	O
insurance	O
fraud	O
.	O
</s>
<s>
Hybrid	O
knowledge/statistical	O
-based	O
systems	O
,	O
where	O
expert	O
knowledge	O
is	O
integrated	O
with	O
statistical	O
power	O
,	O
use	O
a	O
series	O
of	O
data	B-Application
mining	I-Application
techniques	O
for	O
the	O
purpose	O
of	O
detecting	O
cellular	O
clone	O
fraud	O
.	O
</s>
<s>
Link	B-Application
analysis	I-Application
comprehends	O
a	O
different	O
approach	O
.	O
</s>
