Self-service terminal

• US 20090057395A1
• Filed: 09/05/2007
• Published: 03/05/2009
• Est. Priority Date: 09/05/2007
• Status: Active Grant

• Current Assignee: NCR Atleos Corp.
• Original Assignee: NCR Corporation (NCR Voyix Corporation)
• Inventors: Damoulas, Theodoros, Girolami, Mark, He, Chao
• Granted Patent: US 7,942,315 B2
• Time in Patent Office: Days
• US Class Current: 235/379
• CPC Class Codes: G06F 18/2321 using statistics or function optimisation, e.g. modelling of probability density functions
• CPC Class Codes: G06F 18/24155 Bayesian classification
• CPC Class Codes: G06F 18/254 of classification results, e.g. of results related to same input data
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G07F 19/20 Automatic teller machines [ATMs]
• CPC Class Codes: G07F 19/207 Surveillance aspects at ATMs

Abstract

A method of detecting attempted fraud at a self-service terminal. The method involves providing a classifier having a plurality of input zones, each input zone being adapted to receive data of a different modality. The classifier may be a statistical model. The method also involves mapping inputs from a plurality of sensors in the self-service terminal to respective input zones of the classifier. For each input zone, the classifier independently operates on inputs within that zone to create a weighted set; combines the weighted sets using a dedicated weighting function for each weighted set to produce a composite set; and derives a plurality of class values from the composite set to create a classification result. The classification result can be used to predict the probability of abnormal operation, which may be indicative of fraud.

Claims

1. A method of detecting attempted fraud at a self-service terminal, the method comprising:

providing a classifier having a plurality of input zones, each input zone being adapted to receive data of a different modality;

mapping inputs from a plurality of sensors in the self-service terminal to respective input zones of the classifier;

for each input zone, independently operating on inputs within that zone to create a weighted set;

combining the weighted sets using a dedicated weighting function for each weighted set to produce a composite set; and

deriving a plurality of class values from the composite set to create a classification result.

2. A method according to claim 1, wherein operating on inputs within a zone to create a weighted set comprises using base kernels to create a weighted set.

3. A method according to claim 1, wherein deriving a plurality of class values from the composite set to create a classification result includes using regression techniques to derive the plurality of class values.

4. A method according to claim 1, wherein providing a classifier having a plurality of input zones includes providing a Bayesian model as the classifier.

5. A method according to claim 1, wherein the classifier uses different parameters for different operating modes of the self-service terminal.

6. A method of detecting attempted fraud at a self-service terminal, the method comprising:

providing a classifier comprising a statistical model;

mapping inputs from sensors in the self-service terminal into the classifier; and

operating on the inputs using the statistical model to predict a probability of fraud.

7. A method according to claim 6, further comprising communicating an alarm signal to a remote management system in the event that the probability of fraud exceeds a predetermined level.

8. A self-service terminal comprising:

a plurality of customer interaction devices; and

a controller coupled to the customer interaction devices for controlling their operation, the controller including a classifier including a statistical model to predict a probability of fraud based on inputs received from the customer interaction devices.

9. A self-service terminal according to claim 8, wherein the terminal is an automated teller machine (ATM).

10. A method of operating a network of self-service terminals coupled to a management system, where each terminal includes a classifier for predicting fraud based on the operation of that terminal, the method comprising:

receiving a communication from a self-service terminal indicating that the classifier in that terminal has predicted a probability of fraud; and

triggering a dispatch signal to dispatch a customer engineer to the self-service terminal to ascertain if fraud is being perpetrated.

11. A method according to claim 10, wherein the management system is operable to transmit an updated statistical model for a classifier in the event that a new statistical model becomes available.

12. A method according to claim 10, wherein the management system is operable to transmit a risk parameter that indicates a probability of fraud value that must be met for the terminal to indicate fraud.