French Extended Emotional LexiconL
Automatic text analysis to detect the presence of subjective meanings, their polarity (positive, negative and neutral), the associated emotions (joy, anger, fear, etc.), and their intensity has generated numerous studies in the last decade. The applied methods usually depends on the texts nature: tweets (Roberts et al., 2012), mails (Pestian et al., 2012), news headlines (Strapparava and Mihalcea, 2008), etc., and on the application domain: politics (Maynard and Funk, 2012), environment (Hamon et al., 2015), health (Melzi et al., 2014) etc. These methods are generally based on techniques from statistics, Natural Language Processing and Machine Learning. Their efficiency depends mostly on the creation of adapted sentiment lexicons.
To date, most existing affect lexicons have been created for English and for polarity, such as Bing Liu’s Opinion Lexicon (Qiu et al., 2009), the Subjectivity Lexicon (Wilson et al., 2005), etc. However, more specific resources have been created for French such as CASOAR (Asher et al., 2008), and for emotion words such as WordNet Affect (Strapparava and Valitutti, 2004). In this page, we describe the elaboration of a new French lexicon containing more than 14,000 terms according to their polarities and their expressed emotions. The applied method for the elaboration of our new lexicon is based on the translation and expansion to synonyms of the NRC Word-Emotion Association emotion lexicon. An experienced human translator supervised all the automatically translated entries and enriched them. The obtained resource has been evaluated for both classification tasks (polarity and emotion) using various French benchmarks from the literature. Furthermore, the proposed approach is generic and can be adapted to other languages and other resources used as input of the described process.
Last update on 09/10/2015