IBM Classification Module Updated - Classify Unstructured Data

By John Conroy | Published Dec 11, 2007

IBM has announced an upgrade to its Classification Module, which automates the categorization of large volumes of enterprise information, integrating with the FileNet P8 CMS platform.

The system is aimed at data stored in FileNet repositories. It automatically determines the importance of information, then stores and classifies it appropriately.

Large amounts of previously unmanaged content can be processed, as can content already under management but in need of reclassification.

The level of automation is customizable, so if your faith in IBM's algorithms is less than total, you can set the system up with as much oversight as you wish. This is done through configurable “confidence levels” and workflows designed from within the classification review interface.

This review capability uses the Module's real-time learning to give the system feedback to improve accuracy in its classification process.

Cloudmark, an IBM customer, provides anti-spam, anti-phishing and anti-virus solutions for ISPs, mobile operators, enterprises and consumers. Faced with mounting numbers of emails from customers (reaching 500 per day), the cost of processing redundant questions put a significant strain on the company's resources.

Classification Module helped the company discover recurring themes in customer queries and determined which messages were most important, as well as helping with eDiscovery and records management issues for a growing mountain of data.

“Our customers ask complex questions requiring specific information. The IBM Classification Module helps them get to the right answers more quickly,” said Kris Politopoulos, director, customer support, Cloudmark.

“Using the IBM Classification Module, we achieve more than 90 percent accuracy in our automated responses. As a result, we have cut the number of tickets we receive in half and we can effectively support more products without adding support staff. Without the IBM Classification Module, we would have needed to double the size of our team to handle our growing customer base, which would have cost hundreds of thousands of dollars each year.”

View a demo of IBM Classification Module here.