* Machine learning techniques to find fraudulent operations.
* Enabling the fraud analyst to construct rules for fraud detection based on his or her expertise.
The first approach finds fraud faster than the authorization system can, and gives the analyst clues to construct new fraud detection rules. The techniques used fall in the field of artificial intelligence (AI). The latter lets the analyst find fraud by mastering the rules and adapting them to find new fraud patterns. This is known as intelligence amplification (IA). These two approaches reinforce themselves achieving more accuracy in fraud detection over time.
Mebone Fraud enables the possibility
to
catch
fraud in electronic transactions
even
before
the fraud is commited.
Fraud detection techniques range from simple rule engines to Neural Nets (ANN). Mebone Fraud uses Bayesian Networks (BBN). Each technique has owns its pros and cons:
|
||||||||||||||||||||||||||||||
Analyst Know How. Usually, fraud analysts gain a lot of knowledge and understanding of how fraud works from their everyday responsibilities. It is important to be able to leverage this knowledge using the system's ability to understand it. Mebone Fraud allows this knowledge transfer using rules expressed in the analyst's natural language. The system takes care of building syntactically correct rules that avoid misleading and unintentional errors. The Bayesian net is constructed using the analyst knowledge, but suggests to the analyst ways to improve upon the net design. This is not possible using neural nets.
Analysis Performance. At peak moments, an authorization system can manage hundreds of operations per second. Any operation is a potential fraud, so the system must support a very high rate of analysis performance. Both techniques (BBN and ANN) successfully accomplish this objective. In the case of rules, the analysis time grows because of the amount of information being analyzed (real time combined with historical data).
Learning Rate. Automated systems learn from past experience. Both systems (ANN and BBN) need to scan past data to tune their fraud sensitivity. While Neural nets need lots of time (even retraining) Bayesian Networks are much faster as self-tuning.
Feedback. It is important for an automated learning mechanism to explain its decisions. In the case of rules, the explanation is straightforward: one is good if it detects fraud. BBN gives clues about the fraud detection indicating which of the variables used is responsible for an ever changing fraud pattern. Examining the variables can illustrate which of their potential values are increasing fraud. For example if one of the variables is "Country", the BBN indicates not only that the fraud is increasing abroad, it also indicates which country is responsible. ANN keeps this information hidden in its data structure.
Reaction to change. It is crucial to detect fraud pattern as soon as possible. Every technique has its own inertia. This is inevitable, so the system must reduce this inertia as much as possible. The inertia in rules resides in the analyst mind. The inertia in ANN and BBN is inherent to its data structures. BBN nets are much quicker than ANN nets in detecting a new pattern. In some cases the change in pattern is so deep that the ANN must be retrained, consuming lots of human and computer resources.
Analysis Performance. At peak moments, an authorization system can manage hundreds of operations per second. Any operation is a potential fraud, so the system must support a very high rate of analysis performance. Both techniques (BBN and ANN) successfully accomplish this objective. In the case of rules, the analysis time grows because of the amount of information being analyzed (real time combined with historical data).
Learning Rate. Automated systems learn from past experience. Both systems (ANN and BBN) need to scan past data to tune their fraud sensitivity. While Neural nets need lots of time (even retraining) Bayesian Networks are much faster as self-tuning.
Feedback. It is important for an automated learning mechanism to explain its decisions. In the case of rules, the explanation is straightforward: one is good if it detects fraud. BBN gives clues about the fraud detection indicating which of the variables used is responsible for an ever changing fraud pattern. Examining the variables can illustrate which of their potential values are increasing fraud. For example if one of the variables is "Country", the BBN indicates not only that the fraud is increasing abroad, it also indicates which country is responsible. ANN keeps this information hidden in its data structure.
Reaction to change. It is crucial to detect fraud pattern as soon as possible. Every technique has its own inertia. This is inevitable, so the system must reduce this inertia as much as possible. The inertia in rules resides in the analyst mind. The inertia in ANN and BBN is inherent to its data structures. BBN nets are much quicker than ANN nets in detecting a new pattern. In some cases the change in pattern is so deep that the ANN must be retrained, consuming lots of human and computer resources.
The combination of state of the art
techniques, makes Mebone Fraud an
unique tool for fraud detection.
|
Mebone's graphical Interface is simple, and is based on the "Overview and zoom" paradigm. The 'alarms window' presents quickly and clearly all of the information that the analyst needs to investigate the alert. Alarms are prioritized within each window. Lower-priority alarms do not appear in the visible portion of the alarm window. Mebone Fraud users can drill down into the alarm details with a single click. One mouse action brings up an informative table with data relating to the object being analyzed. The analyst can build different rule sets, one for each detection strategy, and all of them can be activated at the same time. Analysts using this tool essentially have a cadre of tireless automated assistants searching for fraud 24 hours a day: Intelligence Amplification. Mebone Fraud provides assistance to rule developers using a text-based grammar-driven real time parser/editor called Rule Builder. Sophisticated real time parsing is used to present all grammatically valid alternatives as the rule is written and all valid alternatives when the rule is being edited. A syntax error cannot be made using this approach. It provides a rich extensible language that can help analysts formulate and execute a wide range of fraud prevention ideas in a natural way. All of the complex information is presented in a graphical view, enabling a quick visual understanding of what is happening, allowing fast, real time decisions. 
 |
|
Graphical Interface is simple, based on
the "Overview and zoom" paradigm.
Mebone offers the information that the
analyst needs in a simple way.
|
Fraud analytics involves many measurements made over an extended period. For example, a good indicator in retail fraud is the average activity of a merchant over the past three months. Without access to historical data, determining this indicator would take many hours, which is clearly unacceptable. Many attempts to use general purpose AI tools for fraud detection are ignoring the large number of details which are making the huge historical data to be accommodated in a real time system. Mebone Fraud includes specialized database engines to accomplish complex historical calculations in seconds over extended periods (e.g. the last three months). Information can be retrieved as long as five years into the past.
With regards to automated fraud detection, the system must run faster than the authorization mechanism. While the authorization makes its own decisions based on card limits, for example, it can and should also launch asynchronous requests to the anti fraud system to get a fraud evaluation for the same transaction. Mebone Fraud can be integrated with SMS brokers to communicate with the cardholder transaction at the moment of purchase.
|
|
Highly optimized algorithms makes
Mebone Fraud capable to analyze card
or electronic transactions in real time.
* Machine learning techniques to find fraudulent operations.
* Enabling the fraud analyst to construct rules for fraud detection based on his or her expertise.
The first approach finds fraud faster than the authorization system can, and gives the analyst clues to construct new fraud detection rules. The techniques used fall in the field of artificial intelligence (AI). The latter lets the analyst find fraud by mastering the rules and adapting them to find new fraud patterns. This is known as intelligence amplification (IA). These two approaches reinforce themselves achieving more accuracy in fraud detection over time.
Fraud detection techniques range from simple rule engines to Neural Nets (ANN). Mebone Fraud uses Bayesian Networks (BBN). Each technique has owns its pros and cons:
|
||||||||||||||||||||||||||||||
Analyst Know How. Usually, fraud analysts gain a lot of knowledge and understanding of how fraud works from their everyday responsibilities. It is important to be able to leverage this knowledge using the system's ability to understand it. Mebone Fraud allows this knowledge transfer using rules expressed in the analyst's natural language. The system takes care of building syntactically correct rules that avoid misleading and unintentional errors. The Bayesian net is constructed using the analyst knowledge, but suggests to the analyst ways to improve upon the net design. This is not possible using neural nets.
Analysis Performance. At peak moments, an authorization system can manage hundreds of operations per second. Any operation is a potential fraud, so the system must support a very high rate of analysis performance. Both techniques (BBN and ANN) successfully accomplish this objective. In the case of rules, the analysis time grows because of the amount of information being analyzed (real time combined with historical data).
Learning Rate. Automated systems learn from past experience. Both systems (ANN and BBN) need to scan past data to tune their fraud sensitivity. While Neural nets need lots of time (even retraining) Bayesian Networks are much faster as self-tuning.
Feedback. It is important for an automated learning mechanism to explain its decisions. In the case of rules, the explanation is straightforward: one is good if it detects fraud. BBN gives clues about the fraud detection indicating which of the variables used is responsible for an ever changing fraud pattern. Examining the variables can illustrate which of their potential values are increasing fraud. For example if one of the variables is "Country", the BBN indicates not only that the fraud is increasing abroad, it also indicates which country is responsible. ANN keeps this information hidden in its data structure.
Reaction to change. It is crucial to detect fraud pattern as soon as possible. Every technique has its own inertia. This is inevitable, so the system must reduce this inertia as much as possible. The inertia in rules resides in the analyst mind. The inertia in ANN and BBN is inherent to its data structures. BBN nets are much quicker than ANN nets in detecting a new pattern. In some cases the change in pattern is so deep that the ANN must be retrained, consuming lots of human and computer resources.
Analysis Performance. At peak moments, an authorization system can manage hundreds of operations per second. Any operation is a potential fraud, so the system must support a very high rate of analysis performance. Both techniques (BBN and ANN) successfully accomplish this objective. In the case of rules, the analysis time grows because of the amount of information being analyzed (real time combined with historical data).
Learning Rate. Automated systems learn from past experience. Both systems (ANN and BBN) need to scan past data to tune their fraud sensitivity. While Neural nets need lots of time (even retraining) Bayesian Networks are much faster as self-tuning.
Feedback. It is important for an automated learning mechanism to explain its decisions. In the case of rules, the explanation is straightforward: one is good if it detects fraud. BBN gives clues about the fraud detection indicating which of the variables used is responsible for an ever changing fraud pattern. Examining the variables can illustrate which of their potential values are increasing fraud. For example if one of the variables is "Country", the BBN indicates not only that the fraud is increasing abroad, it also indicates which country is responsible. ANN keeps this information hidden in its data structure.
Reaction to change. It is crucial to detect fraud pattern as soon as possible. Every technique has its own inertia. This is inevitable, so the system must reduce this inertia as much as possible. The inertia in rules resides in the analyst mind. The inertia in ANN and BBN is inherent to its data structures. BBN nets are much quicker than ANN nets in detecting a new pattern. In some cases the change in pattern is so deep that the ANN must be retrained, consuming lots of human and computer resources.
|
Mebone's graphical Interface is simple, and is based on the "Overview and zoom" paradigm. The 'alarms window' presents quickly and clearly all of the information that the analyst needs to investigate the alert. Alarms are prioritized within each window. Lower-priority alarms do not appear in the visible portion of the alarm window. Mebone Fraud users can drill down into the alarm details with a single click. One mouse action brings up an informative table with data relating to the object being analyzed. The analyst can build different rule sets, one for each detection strategy, and all of them can be activated at the same time. Analysts using this tool essentially have a cadre of tireless automated assistants searching for fraud 24 hours a day: Intelligence Amplification. Mebone Fraud provides assistance to rule developers using a text-based grammar-driven real time parser/editor called Rule Builder. Sophisticated real time parsing is used to present all grammatically valid alternatives as the rule is written and all valid alternatives when the rule is being edited. A syntax error cannot be made using this approach. It provides a rich extensible language that can help analysts formulate and execute a wide range of fraud prevention ideas in a natural way. All of the complex information is presented in a graphical view, enabling a quick visual understanding of what is happening, allowing fast, real time decisions.
|
|
|
Fraud analytics involves many measurements made over an extended period. For example, a good indicator in retail fraud is the average activity of a merchant over the past three months. Without access to historical data, determining this indicator would take many hours, which is clearly unacceptable. Many attempts to use general purpose AI tools for fraud detection are ignoring the large number of details which are making the huge historical data to be accommodated in a real time system. Mebone Fraud includes specialized database engines to accomplish complex historical calculations in seconds over extended periods (e.g. the last three months). Information can be retrieved as long as five years into the past.
With regards to automated fraud detection, the system must run faster than the authorization mechanism. While the authorization makes its own decisions based on card limits, for example, it can and should also launch asynchronous requests to the anti fraud system to get a fraud evaluation for the same transaction. Mebone Fraud can be integrated with SMS brokers to communicate with the cardholder transaction at the moment of purchase.
|
|
Copyright © 2011 Evendor Engineering S.L | Tel: +34915179864 | Fax: +34915179865 | Email: Evendor Engineering