-
Notifications
You must be signed in to change notification settings - Fork 4
ItsdbTreebanking
This page presents user-supplied information, hence may be inaccurate in some details, or not necessarily reflect use patterns anticipated by the [incr tsdb()] developers. This page was initiated by FrancisBond; please feel free to make additions or corrections as you see fit. However, before revising this page, one should be reasonably confident of the information given being correct.
Information here describes the syntax-based discriminants. This mode can be chosen as follows:
(setf lkb::*tree-discriminants-mode* :classic)
The default discriminant mode is :classic, but in the LOGON distribution it defaults to :modern. This is to enable discriminant-based MRS comparison, which is also applicable to parsers other than the LKB or PET (e.g. when using MRS-enabled XLE grammars).
Create a ".tsdbrc" file to set options for treebanking. Dan highly suggests setting the number of analyses and results to 500; 100 is too few and 1000 is no fun. You will need 2Gb+ of memory for this task, so if you don't have this much, find another machine.
(setf tsdb::*tsdb-maximal-number-of-results* 500)
(setf tsdb::*tsdb-maximal-number-of-analyses* 500)
-
- Selecting and rejecting parses
-
- Automatic and Interactive updating to a new grammar
-
- Exporting data from treebanks, as trees, (R)MRSs, dependencies and so on
- Normalizing treebanks (e.g. thinnning)
-
- Creating and scoring stochastic models
If you have an already treebanked corpus (such as the ones that come with the ERG releases), you can look at the gold trees using the [incr tsdb()] graphical interface. In the [incr tsdb()] podium, select the treebank profile and then go to Browse-->Results. To access the gold tree for a particular example, first click on the corresponding number in the "derivation" column. A window will open displaying the tree in the bracketed text format. Click on that text line, and you should see the tree visualization.
Home | Forum | Discussions | Events