Location based authentication system
An improved location based authentication (LBA) system for detecting fraudulent transactions committed by means of misuse of payment cards is disclosed. The improved LBA system performs a series of transaction analysis and generates a fraud-score which provides an indication as whether to authorize an attempted transaction. The system is designed to increase profitability of credit card companies by reducing the accumulated losses due to fraud. In one embodiment, the system detects fraud by analyzing the behavior patterns of card holders and merchants.
1. A method for accurately and adaptively learning the locations of merchants, comprising:
identifying, by an authentication server, pairs of payment transactions which occurred within a predefined time interval and are submitted by different merchants, wherein payment transactions are recorded by a transaction server and the identification occurs as payment transactions are received at the authentication server;
setting, by an authentication server, a maximum allowable velocity and a maximum merchant location score (MLSmax) to their respective initial value;
for each pair of payment transactions, computing, by an authentication server, a transaction velocity;
computing a merchant location score (MLS) for each merchant that submitted a payment transaction in the each pair of payment transactions, if said transaction velocity is lower than a maximum allowable velocity;
setting a location of a merchant of each of the merchants to a location of an anchor location, if said MLS computed for the merchant is above a maximum value of MLS;
saving the computed MLS of a merchant if its computed MLS is below the maximum value of MLS; and
saving, by the authentication server, the location of the merchant in a database, thereby the authentication server adaptively updates the database with the locations of merchants in a specific area as payment transactions are received from the transaction server.
2. The method of claim 1 , further comprising:
setting the location of the merchant according to the location of the anchor location if at least one of the following conditions is satisfied: at least one of said merchants is at said anchor location and said transaction velocity is higher than said maximum allowable velocity.
3. The method of claim 2 , wherein said anchor location is a verified and accurate location.
4. The method of claim 2 , wherein said location and MLS of each merchant are saved in a lookup table.
5. The method of claim 2 , where said location of each merchant is used to determine whether payment transactions should be authorized when performing a location-based analysis.
6. A computer readable medium having stored thereon instructions executable by the computer to perform a process for accurately and adaptively learning the locations of merchants, the process comprising:
identifying pairs of payment transactions which occurred within a predefined time interval and are submitted by different merchants, wherein payment transactions are recorded by a transaction server and the identification occurs as payment transactions are received;
setting a maximum allowable velocity and a maximum merchant location score (MLSmax) to their respective initial value;
for each pair of payment transactions, computing a transaction velocity;
computing a merchant location score (MLS) for each merchant that submitted a payment transaction in the each pair of payment transactions, if said transaction velocity is lower than a maximum allowable velocity;
setting a location of a merchant of each of the merchants to a location of an anchor location, if said MLS computed for the merchant is above a maximum value of MLS;
saving the computed MLS of a merchant if its computed MLS is below a maximum value of MLS; and
saving the location of the merchant in a database, thereby adaptively updating the database with the locations of merchants in a specific area as payment transactions are received.