ESTeR: Combining Word Co-occurrences and Word Associations for Unsupervised Emotion Detection

Findings of the Association for Computational Linguistics: EMNLP 2020, pages 1043–1056 November 16 - 20, 2020. c⃝2020 Association for Computational Linguistics 1043 ESTeR: Combining Word Co-occurrences and Word Associations for Unsupervised Emotion Detection Sujatha Das Gollapalli∗, Polina Rozenshtein∗, See-Kiong Ng Institute of Data Science, National University of Singapore, Singapore {idssdg, idspoli, seekiong}@nus.edu.sg Abstract Accurate detection of emotions in user- generated text was shown to have several ap- plications for e-commerce, public well-being, and disaster management. Currently, the state- of-the-art performance for emotion detection in text is obtained using complex, deep learn- ing models trained on domain-specific, labeled data. In this paper, we propose Emotion- Sensitive TextRank (ESTeR), an unsupervised model for identifying emotions using a novel similarity function based on random walks on graphs. Our model combines large-scale word co-occurrence information with word- associations from lexicons avoiding not only the dependence on labeled datasets, but also an explicit mapping of words to latent spaces used in emotion-enriched word embeddings. Our similarity function can also be computed efficiently. We study a diverse range of datasets including recent tweets related to COVID-19 to illustrate the superior perfor- mance of our model and report insights on pub- lic emotions during the on-going pandemic. 1 Introduction Human beings are known to perceive and feel vari- ous, highly-nuanced emotions, expressed both in spoken and written texts. Modeling emotions in user-generated content has been shown to benefit domains such as commerce, public health, and dis- aster management (Bollen et al., 2011b; Neppalli et al., 2017; Hu et al., 2018; Pamungkas, 2019). E.g., emotion cues from social media posts were used to identify depression and PTSD (Deshpande and Rao, 2017; Arag´on et al., 2019) and for person- alizing chatbots to improve user satisfaction (Wei et al., 2019). Recent studies list as many as 154 human emo- tions (Smith, 2015). However, most researchers in ∗Equal contribution from both authors. Psychology have largely agreed on a set of basic emotions such as anger, fear, disgust, sadness, sur- prise, and happiness (Ekman, 2016) and showed that complex emotions can be expressed using this basic set (Ekman, 1992; Plutchik, 2001). For exam- ple, Plutchik uses combinations, intensity, and op- posites of basic emotions for capturing the higher- order emotions. That is, annoyance and rage can be viewed as the less or more intense forms of anger, and anticipation is the opposite of surprise. Thus, most recent studies on automatic emotion detection use Ekman’s or Plutchik’s sets of 6 or 8 emotions, respectively (Mohammad et al., 2018; Liu et al., 2019). Existing models for automatic emotion identifi- cation in user-generated texts typically use super- vised learning techniques. The state-of-the-art emo- tion detection performance on tweets, news arti- cles, blogs, reviews, and TV-show transcripts is ob- tained using complex, deep learning architectures that combine a range of features including terms, embeddings, and domain-specific aspects such as emojis, as well as human-generated lexicons of emotion-word associations (Chatterjee et al., 2019; Zahiri and Choi, 2018; Mundra et al., 2017; Abdul- Mageed and Ungar, 2017; K¨oper …

Content truncated to 500 words