[jira] [Created] (STANBOL-1263) EntityLinking engines should consider Chunks with NER annotations

["Rupert Westenthaler (JIRA)" <ji...@apache.org>]

Rupert Westenthaler created STANBOL-1263:
--------------------------------------------

             Summary: EntityLinking engines should consider Chunks with NER annotations
                 Key: STANBOL-1263
                 URL: https://issues.apache.org/jira/browse/STANBOL-1263
             Project: Stanbol
          Issue Type: Improvement
          Components: Enhancement Engines
    Affects Versions: 0.12.0
            Reporter: Rupert Westenthaler
            Assignee: Rupert Westenthaler


Detected Named Entities are represented in the AnalyzedText content part by a Chunk with a NER_ANNOTATION. In addition NLP engines may (or may not) also add a PHRASE_ANNOTATION for the same Chunk. However the EntityLinking engines currently only consider PHRASE_ANNOTATION when looking for processable Chunks. Because of that they will not consider Named Entities in cases where NER engines do not provide PHRASE_ANNOTATIONs.

Because especially chunks of NER_ANNOTATIONs are extremely useful for Entity Linking this issue will change the behavior so that Chunks with a NER_ANNOTATION are marked as processable in cases where Nouns are included as processable phrase type in the TextProcessingConfig of the EntityLinkingEngine.

This depends somehow on STANBOL-1262, as without the 'old' processing of Chunks this would result in unintended merging of Noun Phrases and NER chunks.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)