作者: Alexandra Balahur , Jesús M. Hermida , Hristo Tanev
DOI: 10.1007/978-3-642-31782-8_12
关键词:
摘要: In the past years, there has been a growing interest in developing computational methods for affect detection from text. Although much research done field, this task still remains far being solved, as presence of is only very small number cases marked by emotion-related words. rest cases, no such lexical clues emotion are present text and special commonsense knowledge necessary order to interpret meaning situation described understand its affective connotations. light challenges posed emotions contexts which clue present, we proposed implemented base – EmotiNet that stores situations specific felt, represented “action chains”. Following initial evaluations, chapter, describe evaluate two different extend contained EmotiNet: using ontological knowledge. Results show types sources complementary can help improve both precision, well recall implicit systems based on