[[boost_performance]]
Boosting Performance
--------------------
Due to the size of the LVG dictionary, we do not redistribute it in
{osp-short}. You can download it from link:{lvg-home}[] and
swap out the sample dictionary. We have successfully tested the 2008
release in our organization. To use the 2008 release:

. Download either the full version or the lite version from
  http://lexsrv2.nlm.nih.gov/LexSysGroup/Projects/lvg/2008/
. Replace the directory
  +{inst-root-dir}/LVG/resources/lvg/data/HSqlDb+ with +data/HSqlDb+
  from your extracted download.

To update to later versions of LVG:

- Edit the +{inst-root-dir}/LVG/resources/lvg/data/config/lvg.properties+
  file, replacing "lvg2008" with the name of the new release.

To install complete dictionaries for RxNorm, SNOMED-CT, or others that
available through the UMLS, see the following posts on the cTAKES forums
- https://cabig-kc.nci.nih.gov/Vocab/forums/viewtopic.php?f=28&t=423
- https://cabig-kc.nci.nih.gov/Vocab/forums/viewtopic.php?f=28&t=80&start=20#p1459

Models included in {osp-short} may not represent your data
distribution well, if you want to build your own models, please read
<<component_description>>, particularly on:

- <<train_sentdetect_model, training a sentence detector model>>
- <<build_postag_model, building a POS tagger model>>
- <<build_postag_dict, building a POS tag dictionary>>
- <<build_chunker_model, building a chunker model>>
- <<train_dependency_parser, training a dependency parser>>