- From: Said Tabet <stabet@comcast.net>
- Date: Mon, 5 Dec 2005 09:44:23 -0500
- To: <public-rif-wg@w3.org>
- Cc: <Harold.Boley@nrc.ca>, "'Michael Kifer'" <kifer@cs.sunysb.edu>, "'Said Tabet'" <stabet@ruleml.org>, "'Benjamin Grosof'" <bgrosof@MIT.EDU>
Hi everyone, This is a simplified version of a more elaborate use case on fraud detection. Description: ------------ Online fraud is a challenging problem for many companies doing business on the internet. Although many solutions have been implemented, the problem remains a top priority. We considered the fraud detection use case to show how we can achieve the reusability of anti-fraud policies and suport the interoprability and integration requirements (rules, web services, general policies, customer preferences, etc). Subset of rules from the use case Rulebase: ------------------------------------------- <rule id=1>SUSPECT COUNTRY: Orders from a country member of a restricted list with high incidence of fraud require more verification</rule> <rule id=2>SUSPECT EMAIL: Reject the order if email untraceable</rule> <rule id=3>ORDER VELOCITY: If the number of orders per hour exceeds a certain set threshold, flag order and request verification</rule> <rule id=4>MULTIPLE ExPENSIVE ITEMS: If the customer orders multiple items of the same expensive product (exceeding a set limit), flag order and request verification</rule>
Received on Monday, 5 December 2005 14:45:29 UTC