Fraud Detection Rule Optimization
AssigneeBottomline Technologies, Inc.
InventorsDalit Amitai, Shahar Cohen, Yulia Mayer, Avital Serfaty
Life of this patent
Estimating the term…
Abstract
An improved method and apparatus for determining if a financial transaction is fraudulent is described. The apparatus in one embodiment collects transactions off of a rail using promiscuous listening techniques. The method uses linear programming algorithms to tune the rules used for making the determination. The tuning first simulates using historical data and then creates a matrix of the rules that are processed through the linear programming algorithm to solve for the variables in the rules. With the updated rules, a second simulation is performed to view the improvement in the performance. The updated rules are then used to evaluate the transactions for fraud.
Claims20
An improved apparatus for detecting fraud, the apparatus comprising:
- a p-monitor;
- a rail, the p-monitor connected to the rail;
- a data store connected to the p-monitor, the data store containing historical data and a draft rule set;
- wherein the p-monitor creates a matrix of each rule in the draft rule set for each historical transaction in the historical data and solves the matrix for each rule score using linear programming, said rule scores copied into a model rule set;
- and
- wherein the p-monitor monitors the rail for a new transaction, applying data from the new transaction to the model rule set to produce a new transaction score, and if the new transaction score exceeds a threshold, the p-monitor indicates that the new transaction is fraudulent.
A method of detecting a fraudulent transaction on a rail, the method comprising:
- creating a matrix of each rule in a draft rule set for each historical transaction in a historical data stored in a data store;
- solving the matrix for each rule score using linear programming;
- copying said rule scores into a model rule set;
- monitoring the rail for a new transaction;
- applying data from the new transaction to the model rule set;
- producing a new transaction score from the data and the model rule set;
- and
- indicating that the new transaction is fraudulent if the new transaction score exceeds a threshold.
A method of detecting a fraudulent transaction on a rail, the method comprising:
- means for creating a matrix of each rule in a draft rule set for each historical transaction in a historical data stored in a data store;
- solving the matrix for each rule score using linear programming;
- loading said rule scores into a model rule set;
- means for monitoring the rail for a new transaction;
- applying data from the new transaction to the model rule set;
- means for producing a new transaction score from the data and the model rule set;
- and
- means for indicating that the new transaction is fraudulent if the new transaction score exceeds a threshold.
Dependencies are inferred from the claim text.