[jira] [Commented] (CTAKES-231) missing NEs because of inconsistent chunking for parallel sentence constructions

["Sean Finan (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CTAKES-231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15959258#comment-15959258 ] 

Sean Finan commented on CTAKES-231:
-----------------------------------

The dictionary-lookup-fast module has an overlap annotator that can identify all 3 cancers in both example structures.

> missing NEs because of inconsistent chunking for parallel sentence constructions
> --------------------------------------------------------------------------------
>
>                 Key: CTAKES-231
>                 URL: https://issues.apache.org/jira/browse/CTAKES-231
>             Project: cTAKES
>          Issue Type: Bug
>          Components: ctakes-chunker
>    Affects Versions: 3.0-incubating
>            Reporter: James Joseph Masanz
>         Attachments: liver.cancer.chunking.issue.xmi.xml
>
>
> cancer of colon, lung and liver
> results in an annotation for liver cancer
> cancer of colon, liver and lung.
> does not result in an annotation for liver cancer or for lung cancer.
> Thanks Dennis Lee Hon Kit for reporting this.
> Details:
> Reproduced by running 3.0.0-incubating with the separately downloadable UMLS resources, using the AggregatePlaintextUMLSProcessor.xml, results in these chunk annotations:
>  [0] org.apache.ctakes.typesystem.type.syntax.NP
>  [1] org.apache.ctakes.typesystem.type.syntax.PP
>  [2] org.apache.ctakes.typesystem.type.syntax.NP
>  [3] org.apache.ctakes.typesystem.type.syntax.NP
>  [4] org.apache.ctakes.typesystem.type.syntax.PP
>  [5] org.apache.ctakes.typesystem.type.syntax.NP
>  [6] org.apache.ctakes.typesystem.type.syntax.O
>  [7] org.apache.ctakes.typesystem.type.syntax.O
>  [8] org.apache.ctakes.typesystem.type.syntax.NP



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)