Vendor Spam Considerations
May 17, 2010Reputation System Protection
Identifies spam based on reputation of the sender
Accelerated rapidness of zombie IP identification and data distribution to TS servers
Added 15 new spam IP classifiers
Raised effectiveness by 15% in the last month
Message Reputation
Uses reputation of the message content, attachments & meta-data
Added dynamic identification of spam produced by evolving zombies
Raised effectiveness by 13% in the last month
Statistical Message Analysis
Bayesian probability-based analysis of message content
Uses automated training based on Trusted Source feedback
Heuristic & Dictionaries
Created and made available half a dozen new dictionaries for enhanced image spam detection
Correlation Engine
Several other techniques & correlation
www.bestitdocuments.com