http://javelina.cet.middlebury.edu/lsa/out/cover_page.htm
This paper, titled “Patterns in Unstructured Data: Discovery, Aggregation, and Visualization” is offered by the National Institute for Technology and Liberal Education as a “layman’s introduction to latent semantic indexing.” It does start to go into the specifics of how LSI works, but fortunately at an introductory enough level that most of us can grasp. This paper has it all: metadata…ontologies…visualization … what more could you ask for! A helpful introduction to an overlooked technology. See also the NITLE’s LSI project page for pointers to PERL-based LSI software. (Thanks to Bruce Landon for the tip.)- SWL