Hi Sujana, Thanks for writing - SenseClusters is primarily a command line oriented tool, so you can't really use it like a traditional Perl module (where you have methods that you embed in a Perl program). Rather, you would probably use it from the command line, or you could your system ( ) calls within a Perl program (or a bash script, or within any programming language that supports system ( ) )
On to your question - while you could possibly use SenseClusters to compare 1 short text with 2 other ones (to measure text similarity), the intended use of SenseClusters is more so in clustering larger number of short contexts (perhaps 100 or 1000 of them). So, while you could use some of the individual components in SenseClusters to do what you are describing, it would take a bit of digging to get the results back in a form that would be useful for your intended use. The key to doing that would be to represent each of your texts as a short context in the SenseEval-2 format that SenseClusters uses for input. Probably the best source of information about what SenseClusters does and how it does that can be found here : http://search.cpan.org/dist/Text-SenseClusters/Toolkit/README.Toolkit.pod Also, you can get some idea of the intended uses of SenseClusters through the following paper (which doesn't really talk about SenseClusters specifically, but it was definately what I was thinking about when I wrote this... http://www.d.umn.edu/~tpederse/Pubs/pedersen-salr-2007.pdf Computational Approaches to Measuring the Similarity of Short Contexts : A Review of Applications and Methods <http://www.d.umn.edu/~tpederse/Pubs/pedersen-salr-2007.pdf>(Pedersen), to appear in the South Asian Language Review <http://salr.net/>(Also available from CMP-LG E-Print Archive as 0806.3787) <http://arxiv.org/abs/0806.3787> I hope this is of some help. If you remain interested in trying to use SenseClusters for your work I'd be happy to look in a little more detail at what you are trying to do, and figure out how SenseClusters might do that. Cordially, Ted On Mon, May 31, 2010 at 4:11 AM, Mrs sujana jyothi <[email protected]>wrote: > Hi, > > I am Sujana doing my phd at National University of Ireland Maynooth. A part > of my research is text comparision. I am looking at how one text can be > compaired to few other texts and find if the preceeding texts are relevant > to the first one (content analysis). I was looking at LSA which looks at the > similarity between texts (one-to-many). Then I stumbled on > Text::SenseCluster which does take LSA into consideration. > > My query is: Does SenseCluster helps to find if the 2nd, 3rd and so on > texts(paragraphs) are similar to the first text (a paragraph)? > Could you please provide with an example or atleast a link which guides me > how to use this perl module in a perl code... > Does this module already have a corpus which can compare paragraphs? > > I would be grateful if I could get some help in this regard from your end. > I dont think anybody else has used this module as of yet....so I couldnt get > help on the internet. > > Many Thanks. > Regards, > Sujana. > > > > ------------------------------------------------------------------------------ > > > _______________________________________________ > senseclusters-developers mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/senseclusters-developers > > -- Ted Pedersen http://www.d.umn.edu/~tpederse
------------------------------------------------------------------------------
_______________________________________________ senseclusters-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/senseclusters-users
