Brent J. Hecht, Nicole Starosielski, Drew Dara-Abrams
We present a narrative theory-based approach to data mining that generates cohesive stories from a Wikipedia corpus. This approach is based on a data mining-friendly view of narrative derived from narratology, and uses a prototype mining algorithm that implements this view. Our initial test case and focus is that of field-based educational tour narrative generation, for which we have successfully implemented a proof-of-concept system called Minotour. This system operates on a client-server model, in which the server mines a Wikipedia database dump to generate narratives between any two spatial features that have associated Wikipedia articles. The server then delivers those narratives to mobile device clients.
Subjects: 1.10 Information Retrieval; 1.3 Computer-Aided Education
Submitted: Sep 10, 2007