Studying emotions in Romanian words using Formal Concept Analysis. (September 2019)

• Record Type: Journal Article
• Authors: Lupea, Mihaiela; Briciu, Anamaria
• Is Part Of: Computer speech & language. Volume 57(2019)
• Page Start: 128
• Page End: 145
• Publication Date: 2019-09
• Subjects: Natural language processing -- Emotion analysis -- RoEmoLex -- Formal Concept Analysis; Speech processing systems -- Periodicals; Automatic speech recognition -- Periodicals; Computers -- Periodicals; Linguistics -- Periodicals; Speech-Language Pathology -- Periodicals; Traitement automatique de la parole -- Périodiques; Reconnaissance automatique de la parole -- Périodiques; Automatic speech recognition; Speech processing systems; Electronic journals; Periodicals; 006.454
• DOI: 10.1016/j.csl.2019.02.005
• Languages: English
• ISSNs: 0885-2308

Abstract

Highlights: Third version of Romanian Emotion Lexicon (RoEmoLex v.3.) was created. Improvements over RoEmoLex v.2. in the form of revisions and additions. There exists a relationship between emotional content and semantic field of a word. Meaningful word sets are delineated by emotional content in lexicon (EmoSynsets). EmoSynsets can be used to compute relative emotion scores. Abstract: RoEmoLex (Romanian Emotion Lexicon) is a resource developed for emotion detection in Romanian language, now at its third version. Initially translated automatically from an English version and containing terms annotated with eight primary emotions and two polarity tags, this lexicon has undergone a series of changes. New tags were attached to words, including derived emotions, part-of-speech, additional polarity scores and conceptual category information. Moreover, the lexicon was expanded with synonyms of the original terms and words and phrases from similar resources. This paper presents the current version of the emotion lexicon, describing improvements over the previous forms and presenting summary statistics. Through Formal Concept Analysis, a knowledge discovery technique, we examine the structure and content of the lexicon. Dependencies between words at the emotional level are discovered, conceptual hierarchies are built and comparisons with information from another Romanian lexical resource (RoWordNet) are made. The results encourage further research and the possibility to integrate RoEmoLex as a subcomponent of emotion analysis systems for Romanian text.