The Babylon project: toward an extensible text-mining platform

GlaxoSmithKline committed to adopting Verity, an indexing and search engine, to retrieve knowledge within the company intranet. Verity was attractive because it supported taxonomy-oriented browsing and concept-enhanced search. To tap the power in Verity's features, GlaxoSmithKline's Data Exploration Sciences began the Babylon project in late 2002 with two primary goals: develop a text-mining platform that would be the foundation for a variety of text-mining applications and would be extensible to a variety of domains; and develop one or more significant prototype applications on that platform. The first prototype application considered was to mine reports on adverse drug events to identify trends, drug and reaction events, or drug-drug interactions that might not be apparent in nontextual structured databases. The paper discusses the Babylon project.