> Fourth, Bano pointed out that in his MS thesis there is a table (based on > WordNet 1.7) that shows that of the 56,000 compounds known to WordNet, > 56,000 of them only have one sense. Given this, it seems that pos tagging
This is probably a typo, but in case I had misspoken, about 56,000 out of 59,000 compounds in WN1.7 had one sense each. > Fifth, and finally, Bano suggested that we may want to consider using jcn > where we can, and then only using lesk when we need the added ability to > look at different parts of speech. This seems like an excellent > suggestion, and might represent a way to improve performance without > sacrificing quality of results. In general, I think it will be interesting to see whether we can use features of the input pair of senses to choose from a variety of semantic measures. Can we find classes of senses (or sense pairs) that some measures routinely do better at? What features of senses do we need to create these classes? Why are certain measures better at some classes and not at others? Can we make the appropriate choice of measures under different user defined speed/accuracy trade offs? I think there's at least one masters' thesis along these avenues; Ted? :-) I thoroughly enjoyed today's meeting with Jason, Prath, Anagha, Apurva, and you Ted. I hope to do this again in the near future! Thanks for bearing with my questions! Thanks! Bano ------------------------ Yahoo! Groups Sponsor --------------------~--> $9.95 domain names from Yahoo!. Register anything. http://us.click.yahoo.com/J8kdrA/y20IAA/yQLSAA/x3XolB/TM --------------------------------------------------------------------~-> Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/nlpatumd/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
