07-01-2010, 05:19 PM
The use of credit cards has dramatically increased due to development in the e- commerce technology. cases of Credit card fraud are also increasing.
Credit-card-based purchases can be categorized into two
types: 1) physical card
the cardholder presents his card physically
to a merchant for making a payment.
an attacker has to steal
the credit card to show fraud in this case.
2) virtual card.
Here, card number, expiration
date, secure code or similar information are required to make the payment.
The only way to detect
his kind of fraud is to analyze the spending patterns
.
HMM(Hidden markov model) BACKGROUND
An HMM is a double embedded stochastic process with two
hierarchy levels which can be used to model much more
complicated stochastic processes as compared to a traditional
Markov model. An HMM has a finite set of states governed by
a set of transition probabilities. In a given state, an
outcome can be generated according to an
associated probability distribution. only the outcome that is visible to an external observer and not the state . Fields like speech recognition, bioinformatics, and genomics use the HMM. HMM is also used in anomaly detection. HMM was used to model human behavior.
HMM can be characterized by the following:
N is the number of states in the model. the set of states is denoted as S={S1 ; S2 ; . . . SN}
; . . . ; N is an individual state. The state at time
instant t is denoted by qt .
M is the number of distinct observation symbols per
state. The observation symbols correspond to the
physical output of the system being modeled. We
denote the set of symbols V={V1 ; V2 ; . . . VM}
USE OF HMM FOR CREDIT CARD FRAUD DETECTION
A FDS(fraud detection system) runs at a credit card issuing bank. It is sent the card details and the value of purchase to verify
whether the transaction is genuine or not and tries to find any anomaly in the transaction. This calculation is based on spending profile of the cardholder, shipping address, and
billing address, etc.If found to be fraudulent, it raises an alarm, and the issuing bank denies the transaction.
HMM Model for Credit Card Transaction Processing
A credit cardholder makes different kinds of purchases of
different amounts over a period of time.The sequence of types of
purchase is more stable compared to the sequence of
transaction amounts. The set of all possible types of purchase and,
equivalently, the set of all possible lines of business of
merchants forms the set of hidden states of the HMM. the actual items purchased
in the transaction are not determined. After deciding the state and symbol representations, the
next step is to determine the probability matrices A, B, and
mu so that representation of the HMM is complete.
Dynamic Generation of Observation Symbols
we train and maintain an HMM for each cardholder the amount
that the cardholder spent in his transactions are determined from the bank database and K-means clustering algorithm to determine the clusters. The grouping is performed by minimizing the sum of
squares of distances between each data point and the
centroid of the cluster to which it belongs.
Spending profile of cardholders
suggests his normal spending behavior. These are are determined at the
end of the clustering step.
Fraud Detection
After the HMM parameters are learned, we take the
symbols from a cardholderâ„¢s training data and form an
initial sequence of symbols. We input
this sequence to the HMM and compute the probability of
acceptance by the HMM.We input this
new sequence to the HMM formed by dropping O and calculate the probability of
acceptance by the HMM. If O is malicious, the issuing bank
does not approve the transaction, but in other cases is added in the sequence perma-nently, and the new sequence is used as the base sequence for
determining the validity of the next transaction.
Fullseminars report available in the link:
ieeexplore.ieeeiel5/8858/4447479/04358713.pdf
you need IEEE subscription to access it . Your instituition must be having it probably.
Credit-card-based purchases can be categorized into two
types: 1) physical card
the cardholder presents his card physically
to a merchant for making a payment.
an attacker has to steal
the credit card to show fraud in this case.
2) virtual card.
Here, card number, expiration
date, secure code or similar information are required to make the payment.
The only way to detect
his kind of fraud is to analyze the spending patterns
.
HMM(Hidden markov model) BACKGROUND
An HMM is a double embedded stochastic process with two
hierarchy levels which can be used to model much more
complicated stochastic processes as compared to a traditional
Markov model. An HMM has a finite set of states governed by
a set of transition probabilities. In a given state, an
outcome can be generated according to an
associated probability distribution. only the outcome that is visible to an external observer and not the state . Fields like speech recognition, bioinformatics, and genomics use the HMM. HMM is also used in anomaly detection. HMM was used to model human behavior.
HMM can be characterized by the following:
N is the number of states in the model. the set of states is denoted as S={S1 ; S2 ; . . . SN}
; . . . ; N is an individual state. The state at time
instant t is denoted by qt .
M is the number of distinct observation symbols per
state. The observation symbols correspond to the
physical output of the system being modeled. We
denote the set of symbols V={V1 ; V2 ; . . . VM}
USE OF HMM FOR CREDIT CARD FRAUD DETECTION
A FDS(fraud detection system) runs at a credit card issuing bank. It is sent the card details and the value of purchase to verify
whether the transaction is genuine or not and tries to find any anomaly in the transaction. This calculation is based on spending profile of the cardholder, shipping address, and
billing address, etc.If found to be fraudulent, it raises an alarm, and the issuing bank denies the transaction.
HMM Model for Credit Card Transaction Processing
A credit cardholder makes different kinds of purchases of
different amounts over a period of time.The sequence of types of
purchase is more stable compared to the sequence of
transaction amounts. The set of all possible types of purchase and,
equivalently, the set of all possible lines of business of
merchants forms the set of hidden states of the HMM. the actual items purchased
in the transaction are not determined. After deciding the state and symbol representations, the
next step is to determine the probability matrices A, B, and
mu so that representation of the HMM is complete.
Dynamic Generation of Observation Symbols
we train and maintain an HMM for each cardholder the amount
that the cardholder spent in his transactions are determined from the bank database and K-means clustering algorithm to determine the clusters. The grouping is performed by minimizing the sum of
squares of distances between each data point and the
centroid of the cluster to which it belongs.
Spending profile of cardholders
suggests his normal spending behavior. These are are determined at the
end of the clustering step.
Fraud Detection
After the HMM parameters are learned, we take the
symbols from a cardholderâ„¢s training data and form an
initial sequence of symbols. We input
this sequence to the HMM and compute the probability of
acceptance by the HMM.We input this
new sequence to the HMM formed by dropping O and calculate the probability of
acceptance by the HMM. If O is malicious, the issuing bank
does not approve the transaction, but in other cases is added in the sequence perma-nently, and the new sequence is used as the base sequence for
determining the validity of the next transaction.
Fullseminars report available in the link:
ieeexplore.ieeeiel5/8858/4447479/04358713.pdf
you need IEEE subscription to access it . Your instituition must be having it probably.
