Thanks again Ted! This gives me a lot to explore. So far, I've had some reasonable success applying POS to presidential speeches, which you can see in the first paper here: http://geocities.com/pmuhl78/
I'll try out sense tagging in a later iteration of this project & let the list know if this reveals new insights. Cheers, Peter On Mon, Dec 8, 2008 at 8:56 PM, Ted Pedersen <[email protected]> wrote: > Hi Peter, > > Glad to be of some help. SenseRelate::AllWords takes part of speech > tagged input pretty much as it comes out of the Brill Tagger, for > example. I think if you really need to assign meanings to words, then > SenseRelate could be helpful. If you simply need to make a distinction > between "line" as a noun versus "line" as a verb for your > applications, then perhaps POS tagging is enough. POS tagging is > relatively accurate, so you probably want to try that first and see > where it gets you. > > In terms of quick ways to do relation extraction, you might want to > think in terms of using a chunker of some kind to identify noun > phrases and verb phrases, and then use the verb phrases as relations > between noun phrases....some things having to do with relation > extraction can be approximately fairly simply with syntax of this > sort, so you might want to look into that too. I don't have quite as > much experience with chunkers, etc. so don't have any specific > recommendations, but perhaps others who are on this list do... > > Good luck, and keep us posted on how it goes. > > Cordially, > Ted >
