You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@opennlp.apache.org by "Jörn Kottmann (JIRA)" <ji...@apache.org> on 2011/07/19 10:43:58 UTC
[jira] [Closed] (OPENNLP-230) Name Finder should calculate probs
for spans differently
[ https://issues.apache.org/jira/browse/OPENNLP-230?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jörn Kottmann closed OPENNLP-230.
---------------------------------
Resolution: Fixed
Assignee: Jörn Kottmann
> Name Finder should calculate probs for spans differently
> --------------------------------------------------------
>
> Key: OPENNLP-230
> URL: https://issues.apache.org/jira/browse/OPENNLP-230
> Project: OpenNLP
> Issue Type: Bug
> Components: Name Finder
> Affects Versions: tools-1.5.1-incubating
> Reporter: Jörn Kottmann
> Assignee: Jörn Kottmann
> Priority: Minor
> Fix For: tools-1.5.2-incubating
>
>
> The name finder has a method to calculate the probability of a named entity. The method also supports named entites which span multiple tokens. The name finder does sequence labeling, and has a prob per token. To compute the prob of a name span, it simply multiplies the probs of the tokens covered by the span.
> Instead of multiplying it should calculate the arithmetic mean of the token probabilities.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira