Case Study: KPN Research builds a portable web page recommender with Franz Inc.'s Common Lisp and Allegro tools

KPN Research builds a recommender system in 6 months with Franz Inc.

KPN Research, the research department of the largest Dutch telecom operator, faced the challenge of creating a new broadband service concept. Their goal was to develop a personalized, easy-to-use browser interface for a mobile webpad that could provide intelligent website recommendations based on a user's interests and surfing patterns. They engaged Franz Inc. to build a high-performance recommender system that could process millions of URLs.

Franz Inc. implemented a solution using Common Lisp with Allegro CL, AllegroStore, and AllegroServe. This allowed a small team to develop a powerful and unique recommendation engine within a six-month timeframe, a feat they believed would have been too slow with Java or too complex with C++. The resulting system successfully combined collaborative filtering with advanced document classification and was integrated into a user-friendly personal browser. The software was successfully piloted and is now being reviewed by publishers for knowledge management and content distribution applications.

KPN Research