Analysis on Credit Card Fraud Detection

Methods
ABSTRACT:
Due to the rise and rapid growth of E-Commerce, use of credit cards for online purchases
has dramatically increased and it caused an explosion in the credit card fraud. As credit
card becomes the most popular mode of payment for both online as well as regular
purchase, cases of fraud associated with it are also rising. In real life, fraudulent
transactions are scattered with genuine transactions and simple pattern matching
techniques are not often sufficient to detect those frauds accurately. Implementation of
efficient fraud detection systems has thus become imperatie for all credit card issuing
ban!s to minimi"e their losses. #any modern techniques based on Artificial Intelligence,
Data mining, $u""y logic, #achine learning, %equence Alignment, &enetic 'rogramming
etc., has eoled in detecting arious credit card fraudulent transactions. A clear
understanding on all these approaches will certainly lead to an efficient credit card fraud
detection system. (his paper presents a surey of arious techniques used in credit card
fraud detection mechanisms and ealuates each methodology based on certain design
criteria.
EXISTING SYSTEM
(he (raditional detection method mainly depends on database system and the
education of customers, which usually are delayed, inaccurate and not in-time.
After that methods based on discriminate analysis and regression analysis are
widely used which can detect fraud by credit rate for cardholders and credit card
transaction.
$or a large amount of data it is not efficient.
R!B"EM REC!GNITI!N
(he high amount of losses due to fraud and the awareness of the relation between
loss and the aailable limit hae to be reduced.
(he fraud has to be deducted in real time and the number of false alert has to be
minimi"ed.
R!!SED SYSTEM
(he proposed system oercomes the aboe mentioned issue in an efficient way.
)sing genetic algorithm the fraud is detected and the false alert is minimi"ed and it
produces an optimi"ed result.
(he fraud is detected based on the customer*s behaior. A new classification
problem which has a ariable misclassification cost is introduced.
+ere the genetic algorithms is made where a set of interal alued parameters are
optimi"ed.

SYSTEM ARC#ITECT$RE
DA(A
,A-E+.)%E
/C)%(.#E-
DA(A0
DA(A
,A-E+.)%E
/C)%(.#E-
DA(A0
-)1E% E2&I2E
-)1E% E2&I2E
&E2E(IC
A1&.-I(+#
&E2E(IC
A1&.-I(+#
$I1(E- 3
'-I.-I(4
$I1(E- 3
'-I.-I(4
$-A)D -)1E
%E(
$-A)D -)1E
%E(
#ARD%ARE RE&$IREMENTS
5 %4%(E# 6 'entium I7 8.9 &+"
5 +A-D DI%: 6 9; &<
5 #.2I(.- 6 => 7&A colour
5 #.)%E 6 1ogitech.
5 -A# 6 8>? #<
5 :E4<.A-D 6 ==; !eys enhanced.
S!FT%ARE RE&$IREMENTS
5 .perating system 6 ,indows @' 'rofessional
5 $ront End 6 AA7A
5 (ool 6 2et<eans IDE
M!D$"ES
)ser &)I
Critical 7alue Identification
$raud Detection using &enetic Algorithm
M!D$"ES DESCRITI!N
$ser G$I
In this module, )ser Interface module is deeloped using Applet 7iewer. (his
module is deeloped to user to identify the credit card fraud using genetic algorithm
technique. %o the user interface must be capable of proiding the user to upload the
dataset and ma!e manipulations and finally must show the user whether fraud has been
detected or not. .nly final output will be in applet screen. All the generation details
/crossoer and mutation0 will b in the console screen of eclipse.
Critical 'alue Identi(ication
Based on CC usa)e Fre*uency
float ccfreq B$loat.alue.f/tempCDE0F$loat.alue.f/tempC?E0G
if/ccfreqH;.80
I
if/$loat.alue.f/tempCJE0H/>Kccfreq00
I
resC;EB=G
resC=EB/$loat.alue.f/tempCJE0Kccfreq0G
L
L
if/resC;EM=0
I
resC=EB/float0ccfreqG
L
Ccfreq B (otal number card used /C)0 F CC age
If ccfreq is less than ;.8 , it means this property is not applicable for fraud and
critical alue Bccfreq
.therwise, it chec! for condition of fraud /i.e0 B
$raud condition B number of time Card used (oday /C)(0 H/ > K ccfreq0
If true, there may chance for fraud using this property and its critical alue is C)(Kccfreq
If flase, no fraud occurance and critical alue Bccfreq
Based on CC usa)e "ocation
int locBInteger.alue.f/tempCNE0G
if//locMB >0 33 /Integer.alue.f/tempCOE0H/ 8 K loc000
I
resC;EB=G
resC=EB/$loat.alue.f/loc0F $loat.alue.f/tempCOE00G
L
if/resC;EM=0
I
resC=EB/float0;.;=G
L
2umber of locations CC used so far /loc0 obtained from dataset/loc0
If loc is less than >, it means this property is not applicable for fraud and critical alue
B;.;=
.therwise, it chec! for condition of fraud /i.e0 B
$raud condition B number of locations Card used (oday /C)(0 H/ > K loc0
If true, there may chance for fraud using this property and its critical alue is locFC)(
If flase, no fraud occurance and critical alue B;.;=
Based on CC !+erDra(t
float od B$loat.alue.f/tempC>E0F$loat.alue.f/tempCDE0G
if/odMB;.80
I
if/$loat.alue.f/tempC=;E0HB=0
I
resC;EB=G
resC=EB/$loat.alue.f/tempC=;E0Kod0G
L
L
if/resC;EM=0
I
resC=EB/float0odG
L
2umber of times CC oerdraft w.r.t C) occurred so far /od0 can be found as,
.d w.r.t C) B .DFC)
If .d w.r.t C) is less than ;.;8, it means this property is not applicable for fraud and
critical alue B .d w.r.t C)
.therwise, it chec! for condition of fraud /i.e0 B
$raud condition B chec! whether oerdraft condition occurred today from /.D( dataset0
If true, there may chance for fraud using this property and its critical alue is .D( K .d
w.r.t C)
If flase, no fraud occurance and critical alue B .d w.r.t C)
Based on CC Boo, Balance
float bb B$loat.alue.f/tempC8E0F$loat.alue.f/tempC9E0G
if/bbMB;.8>0
I
resC;EB=G
resC=EB/$loat.alue.f/80Kbb0G
L
if/resC;EM=0
I
resC=EB/float0bbG
L
%tandard <oo! balance can be found as,
<b B current << F Ag. <<
If bb is less or equals than ;.8>, it means this property is not applicable for fraud and
critical alue B bb
.therwise, it chec! for condition of fraud /i.e0 B
If true, there may chance for fraud using this property and its critical alue is curr<< K bb
If flase, no fraud occurance and critical alue B bb
Based on CC A+era)e Daily S-endin)
float monB $loat.alue.f/tempC?E0FD;G
float balB =;;;;; - $loat.alue.f/tempC9E0G
float tot B monKbalG
float ds BtotF$loat.alue.f/tempC?E0G
if//=;Kds0M$loat.alue.f/tempC==E00
I
resC;EB=G
if/$loat.alue.f/tempC==E0H;0
resC=EB/$loat.alue.f/tempC==E0F /=;Kds00G
else
resC=EB/float0 ;.;G
L
if/resC;EM=0
I
resC=EB/float0;.;=G
L
Fraud Detection usin) Genetic Al)orith.
In this module the system must detect whether any fraud has been occurred in the
transaction or not. It must also display the user about the result. It is calculated based on
following6
Age of CC in months can be calculated using CCage /from dataset0 by,
Age of cc by month B CCageFD;
(otal money being spent from the aailable limit /= la!h P =;;;;;0
<al B =;;;;; Q ag <<
%o, total money spent can be found as,
(ot B Age of cc by month K <al
(otal money spent on each month can be calculated as,
DsBtotK Age of cc by month
it chec! for condition of fraud /i.e0 B
$raud condition B /=; K ds0 is amount spent today /Amt( in dataset0
If true, there may chance for fraud using this property and its critical alue is Amt(F
/=;Kds0
If flase, no fraud occurance and critical alue ;.;=
REFERENCE:
%.<enson Edwin -aR, A. Annie 'ortia, SAnalysis on Credit Card $raud Detection
#ethodsT, IEEE International Conference on Computer, Communication and Electrical
(echnology, IEEE March /0112