Hi Prem,
On Feb 9, 2013, at 2:01 AM, Richard Eckart de Castilho <eckart@ukp.informatik.tu-darmstadt.de>
wrote:
> Am 09.02.2013 um 01:38 schrieb Prem Devanbu <devanbu@cs.ucdavis.edu>:
>
>> We want to annotated some text with a syntax tree, and we're having trouble finding
examples or tutorial content on how this should be done within UIMA. Any pointers appreciated.
You can also represent a syntactic tree simply as offsets with tag (assuming a constituent
structure tree without overlaps), if the hierarchy is implicit in the offset span. Potential
matching offsets have to be handled by other rules (e.g. what cat/tag dominates what other
can be handled independently in subsequent components, or a final visualization). Maybe a
look at the stand-off annotation approaches in for example TEI XML would help.
> The typical way to represent a tree is defining a tree node type like this:
>
> Node {
> Node parent;
> FSArray<Node> children;
> String tag;
> }
>
> An example for such a type is the "Constituent" type in the DKPro Core Syntax type system:
> (...)
--
Dr. Damir Cavar
http://cavar.me/damir/
FaceTime: dcavar@me.com
|