SpotPlug: Open Source real time fraud detection - Presentation Transcript
SpotPlug Detect fraud without paying a cent
Features
Plug and play: it just works without customizations
Handles high volumes and rates of transactions
Very customizable, can write your own rules
Based on Drools Fusion: a powerful CEP engine
User friendly
Open Source: ASL license
Features (continued)
Exploit the possible fraud database
Various connectors
JMS, Web services and Sockets
Databases
Files
SaaS version available
Vertical industries available
Or generic anomalous behavior
Architecture
Fraud patterns and probability
There are 11 patterns supported and implemented.
Some add probability of a fraud, others are more imperative and reject transactions.
This behavior is configurable.
We suggest connecting Streamline to your transaction manager and let stream run through it during an evaluation phase. Then analyze Streamline's output to understand its behavior on your environment and business context.
Spot Plug is an open source real-time fraud detecti more
Spot Plug is an open source real-time fraud detection tool. It is based on the Drools open source project. It's designed to get event stream from different types of sources such as JMS queues, databases and web services.
Spot Plug detects frauds using a variety of built-in real-time patterns. Of course you can add new ones or customize the existing ones to suit your needs. You don't need any other tool, because the patterns are expressed as DRL files (the default Drools rule description file).
It has a standard Swing GUI for following the performance of the rules in real-time. You can analyze the events that match the fraud patterns and its attributes for optimizing the tool detection capabilities.
Spot Plug stores fraud data in a standard JDBC database (PostgreSQL recommended for maximum performance). You can analyze this data using regular data mining tools such as Weka to find similarities and further refine the rules to catch the most common kinds of fraud. less
0 comments
Post a comment