I think there were some talks about this on the conference I went to recently. Keywords might be "natural language processing". Linked is the abstracts of the conference, which you might find some use in.
http://www.insna.org/PDF/Sunbelt/4_ProgramPDF.pdf One alternative I briefly considered is to use google's suggest feature. Two examples: say you want to know if 'reel' and 'fish' belong together, you can try http://google.com/complete/search?q=fishing+r If you want to know whether 'games' and 'cow' belong together, or 'games' and 'herring' http://google.com/complete/search?q=games+herring But it doesn't work that great. Wordnet seems like a decent start, but if you run into better databases for relations between words, let us know :-). For the more practical part. I found displaying large graphs well doable using a combination of java's processing library (Incanter has an extension for that), and a spring vertlets library called Toxiclib for finding a good way to display the network. Personally, I found that an algorithm that works well was one that first used constrained- springs (dampened springs) between nodes of the network, and once that had settled into relative peace, used mindistance-springs to get overlapping nodes separated from each other. http://processing.org/discourse/yabb2/YaBB.pl?num=1238682023 http://toxiclibs.org/ -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
