DepecheMood++: A Bilingual Emotion Lexicon Built Through Simple Yet Powerful Techniques

Abstract

Several lexica for sentiment analysis have been developed and made available in the NLP community. While most of these come with word polarity annotations (e.g. positive/negative), attempts at building lexica for finer-grained emotion analysis (e.g. happiness, sadness) have recently attracted significant attention. Such lexica are often exploited as a building block in the process of developing learning models for which emotion recognition is needed, and/or used as baselines to which compare the performance of the models. In this work, we contribute two new resources to the community: a) an extension of an existing and widely used emotion lexicon for English; and b) a novel version of the lexicon targeting Italian. Furthermore, we show how simple techniques can be used, both in supervised and unsupervised experimental settings, to boost performances on datasets and tasks of varying degree of domain-specificity.

Committed to Health: Key Factors to Improve Users’ Online Engagement through Facebook

“…The index measures the users' engagement through interactive behaviors: number of likes, comments and shares of a post, which define the three main dimensions of the indicator: popularity (P), commitment (C) and virality (V). Different software programs are available for social-media data gathering and quantitative analysis; others are designed to analyze the data from a linguistic point of view, specifically regarding content and emotions [62]. Despite the existence of these, the authors and their research group have developed an ad hoc tool (called "Facebook data model") for this type of research, based on Microsoft technology and previously tested and used in other studies [45,63].…”

“…Finally, and to a lesser extent, HOs echo news related to political, social and environmental issues that may affect the health sector (8.76%) on their Facebook pages. As [62] point out, the linguistic usage is an interesting characteristic to consider in the context of social media, and many sentiment lexica have been developed to classify post messages, generally as positive and negative. In this sense, [14] state that Facebook posts are generally positive.…”

“…In this line, our results show that, since 59.62% of the posts show positive expressions, the remaining 40.38% is distributed almost equally between expressions that show negative feelings or are made with expressions of a neutral nature (Figure 4). As [62] point out, the linguistic usage is an interesting characteristic to consider in the context of social media, and many sentiment lexica have been developed to classify post messages, generally as positive and negative. In this sense, [14] state that Facebook posts are generally positive.…”

MoralStrength: Exploiting a moral lexicon and embedding similarity for moral foundations prediction

“…With time, not only the methods became more sophisticated but also the tasks become more ambitious. Extensive studies on linguistic markers of sentiment and affect [19,20,17,21,22] paved the way to assess more complex constructs such as personality [23,24] and human values [25,26]. Moral values are considered to be a higher level construct with respect to personality traits, determining how and when dispositions and attitudes relate with our life stories and narratives [27].…”

EMOMINER at SemEval-2019 Task 3: A Stacked BiLSTM Architecture for Contextual Emotion Detection in Text

“…We exploit the DepecheMood affective lexicon Deepechemood++ (Araque et al, 2018) that has been built in a completely unsupervised fashion, from affective scores assigned by readers to news articles. DepecheMood++ allows for both high-coverage and high-precision, providing scores for 187k entries on the following affective dimensions: Afraid, Happy, Angry, Sad, Inspired, Don't Care, Inspired, Amused, Annoyed.…”