Training Emotion Detector with Emotion Lexicon

[luigi-asprino]

Depends on

Independent

Description

The emotion detector uses lexica annotated with emotions for strengthening its capabilities in recognising emotions in users' sentences.

Source

WP3-WP6 interpackage working group.

Competency Questions

• What are the emotions associated with a given word?
• What are the emotions associated with a particular sense of a certain word?

Ontology/Ontology Design Patterns Relevant for the scenario

• WordNet schema might be considered to specify lexicon
• Framester's DepecheMood graph

SPARQL Questions

The following query returns the emotion score associated with wordnet's synsets.

PREFIX fschema: <https://w3id.org/framester/schema/>
PREFIX depmood: <https://w3id.org/framester/depechemood/depechemood2wn/>
PREFIX wn30schema: <https://w3id.org/framester/wn/wn30/schema/>

SELECT DISTINCT ?syn  (str(?afraid) AS ?Afraid) (str(?amused) AS ?Amused) (str(?angry) AS ?Angry) (str(?annoyed) AS ?Annoyed) (str(?dontcare) AS ?DontCare) (str(?happy) AS ?Happy) (str(?sad) AS ?Sad)
WHERE{
  ?syn depmood:AFRAIDscore  ?afraid ;
    depmood:AMUSEDscore ?amused ;
    depmood:ANGRYscore ?angry ;
    depmood:ANNOYEDscore ?annoyed ;
    depmood:DONT_CAREscore ?dontcare ;
    depmood:HAPPYscore  ?happy ;
    depmood:INSPIREDscore   ?insipered ;
    depmood:SADscore   ?sad .
}

The following query retrieves words and glosses associated with sense synsets.

SELECT * {
  ?syn wn30schema:containsWordSense ?ws .
  ?syn wn30schema:gloss ?gloss .
  ?ws wn30schema:word ?w .
  ?w wn30schema:lexicalForm ?lf .
}

Toy Dataset/SPARQL Endpoint

Framester's Endpoint