Rupert Westenthaler created STANBOL-1030:
--------------------------------------------
Summary: EntityLinking should sort Suggestions with the same score
based on entity rankings
Key: STANBOL-1030
URL: https://issues.apache.org/jira/browse/STANBOL-1030
Project: Stanbol
Issue Type: Improvement
Components: Enhancement Engines
Affects Versions: enhancement-engines-0.10.0
Reporter: Rupert Westenthaler
Assignee: Rupert Westenthaler
Priority: Minor
Vocabularies the EntityLinkingEngine links against can define EntityRanking. An
entity ranking defines the popularity (similar to page-rank) of an Entity
within the knowledge base. For Wikipedia this is e.g. generated by the number
of incoming links from other wikipedia pages.
While the EntityRanking is defined in the internal API of the
EntityLinkingEngine it is currently not used for the score of linked entities.
This is mainly because the score represent how well the label of an Entity does
match the section in the processed text.
This will introduce a new option that allows for Suggestions that would have
the same score (typically 1.0 - for exact matches) to get their score slightly
(less than 0.1) changed so that they are ranked based on their entity ranking.
For linking against wikipedia this will ensure that Paris (France) will have a
slightly higher ranking as Paris (Texas, US).
Those changes MUST NOT change the order of suggestions for entities that do not
have the same score.
Users will be able to enable/disable this feature via a setting of for the
entity linking configuration. By default this feature will be activated.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
