ISL: Carving Into Web-Logs for Business Intelligence
ACTION ITEMS

ISL's Clementine Data Mining System is to be used as the basis for a new tool to analyse web-log data, aiming to revolutionise Internet commerce. The tool, which will be developed as part of the MIMIC project (Mining the Internet for Marketing Intelligence), will be a web-enabled data mining system, customised to deal with web-log data.

MIMIC is a new collaborative venture between Clementine suppliers ISL, the University of Ulster, and web-site providers Adnet from Dublin (Ireland), and At Internet from Bordeaux (France). "We want to bring the subtlety and power of modern marketing technology into the exciting new domain of the World Wide Web," says Maurice Mulvenna of the University of Ulster.

To supplement the data mining functionality already available in Clementine, new facilities for automatically decoding web-log data will be developed. The new tool will also be web-enabled, allowing long-distance data mining via an Internet browser. The route a user takes when browsing a web-site is logged and stored. Increasingly, users can also be tracked over repeated visits via Internet cookies.

By analyzing this information with data mining techniques, it is hoped that hosts will be able to improve the design and functionality of their web sites. "To date, we don't have a very adequate data model for web-log information," said Alan Montgomery of ISL. "With MIMIC, one of the most important objectives is to discover the kind of information that is available, and what it can tell us about how web-sites are used."

"MIMIC will open us the way to one to one marketing," said Louis Moussy of At Internet. At Internet was founded in January 96 and registered in Bordeaux, France. As an ISP, Electronic Commerce became prominent in its activities. Its malls include more than 50 first ranked "chateaux", and many gourmet products vendors. At Internet is in the leading team of the most visited French servers. Trust is of paramount importance in the EC domain, inclining At Internet to present the most adapted products to client profiles.

The University of Ulster employs over 2500 staff and has 20,000 FTE students, studying at the four campuses of the University in Magee College (Derry), Jordanstown (Belfast) Coleraine and the Art College in Belfast. The University will soon have a fifth campus, at Springvale in West Belfast. Researchers in NIKEL, part of the University's Faculty of Informatics, will carry out research work to enable the Internet specific features, and to develop new algorithms.

ISL is a specialist data-mining company based in the UK, developing and supplying the Clementine Data Mining System. Clementine is a premier interactive data mining toolkit. Formed in 1989, ISL began as a management buy-out of the AI Products Divisions from SD-SCICON (later to become EDS). From 1990 onwards, ISL staff worked on many different data mining projects for customers. In 1992, ISL's Colin Shearer initiated Project Clementine, aiming to deliver an integrated data mining system, accessible to data owners. The first commercial release came in 1994, and Clementine Version 5.0 is due in Autumn 1998. ISL now has 550 customers in 38 countries, including Reuters, Unilever, Singapore Telecoms and the BBC. ISL Decision Systems Inc., based near Philadelphia, was opened in 1996.

Data Mining and Clementine

Business people tend to develop an intuitive feel for their line of work, based on their experience. Data mining allows this kind of learning from experience to be done computationally, meaning that hunches can be investigated much more efficiently, directly on the evidence of the information available.

Although it incorporates statistics, data visualisation and machine intelligence, data mining is essentially a human process. By looking at the patterns and trends, and building models of features in the data, data miners can get to know the detail of their business, quickly and accurately. Yet until recently, serious data analysis was only accessible to technical and statistical experts.

The Clementine Data Mining System enables business users to mine their own data. By allowing the people who need the answers to find them, ISL has found that data mining becomes more efficient, more creative and more profitable.