Carrot2 1 document to 1 Cluster only

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Carrot2 1 document to 1 Cluster only

willycws
This post has NOT been accepted by the mailing list yet.
Hi,

Is it possible to have the Carrot2 using Lingo Algorithm to classify document to a single cluster instead of multiple cluster of related documents? Is there a procedure to call with certain parameters to do this?

Willy
Reply | Threaded
Open this post in threaded view
|

Re: Carrot2 1 document to 1 Cluster only

Stanislaw Osinski
Administrator
Hi,

Apologies for a delayed reply, if you don't subscribe to the mailing list when posting, the messages don't reach us by e-mail, so it's easy to miss them.

Carrot2 algorithms create overlapping clusters by design, so it's not easy to modify them to deliver exclusive clusterings.

One solution might be to post-process the results and remove documents from overlapping clusters based on some criteria (e.g. keep the document only in the highest-scoring cluster). After such pruning, you may end up with one- or zero-document clusters, which you may want to remove too.

Thanks!

Staszek

willycws wrote
Hi,

Is it possible to have the Carrot2 using Lingo Algorithm to classify document to a single cluster instead of multiple cluster of related documents? Is there a procedure to call with certain parameters to do this?

Willy