i'm not sure if this is a well-posed question as (unless one opts for
something objective, and non-computable, like kolmogorov complexity) any
metric will be subjective in some way or other. as an example, string
edit distance could be used as your metric (sentences A and B are equally
similar to C if A and B can be mapped to C using the same number of
edits), but one could easily imagine another set of editing operations
that would say that (for example) A could be mapped to C using fewer
operations.
now, if you have some sort of application in mind, then it might be
possible to invent some metric that gave useful similarity measures
between sentences.
do you?
Miles Osborne
On Mon, 13 Nov 2000, Constantin Orasan wrote:
> Hello everybody.
>
> I would like to compute the similarity between two sentences. Could you
> indicate some work which proposes measures for this? I am particularly
> interested in methods which use, in addition to the words, some
> linguistic information attached to the words (e.g. PoS tags, WordNet
> senses, etc.).
>
> Thank you,
> Constantin
>
> Computational Linguistics Group
> University of Wolverhampton
> http://www.wlv.ac.uk/~in6093
>
>
This archive was generated by hypermail 2b29 : Mon Nov 13 2000 - 16:48:03 MET