This talk will present the results of a series of research projects that have the common goal of building concept hierarchies automatically from text. I will describe the basic means of building the hierarchies, tests of their properties and user tests of their utility in aiding searchers to find relevant items. I will also describe the range of collections that the hierarchies have been tested on and why some collections work substantially better than others. I will also outline our recent thoughts on what is making the hierarchies so liked by users.
Mark Sandrson is a senior lecturer at the University of Sheffield. His research is in information retrieval (IR) (e.g. web search engines), cross language IR (CLIR), summarisation, image retrieval by captions, word sense ambiguity, and anything else to do with unstructured data.