Exploring pathways across stories

This paper describes a method for supporting the exploration of a collection of documents organized as a hypertext by investigating relations between documents along user-specified paths. The approach is demonstrated on a corpus of stories about the World War Two activities of the British Government Code and Cypher School at Bletchley Park. Each story is described by one or more events and annotated in terms of domain ontologies. A pathway in the document space is a sequence of events in which adjacent events share common binding concepts. The criteria for selecting the pathway include a measure of the adherence to the user-specified part of the document space and the mutual information between adjacent documents calculated from their annotations.