SenWave: Monitoring the Global Sentiments under the COVID-19 Pandemic

Abstract

Since the first alert launched by the World Health Organization (5 January, 2020), COVID-19 has been spreading out to over 180 countries and territories. As of June 18, 2020, in total, there are now over 8,400,000 cases and over 450,000 related deaths. This causes massive losses in the economy and jobs globally and confining about 58% of the global population. In this paper, we introduce SenWave, a novel sentimental analysis work using 105+ million collected tweets and Weibo messages to evaluate the global rise and falls of sentiments during the COVID-19 pandemic. To make a fine-grained analysis on the feeling when we face this global health crisis, we annotate 10K tweets in English and 10K tweets in Arabic in 10 categories, including optimistic, thankful, empathetic, pessimistic, anxious, sad, annoyed, denial, official report, and joking. We then utilize an integrated transformer framework, called simpletransformer, to conduct multi-label sentimental classification by fine-tuning the pre-trained language model on the labeled data. Meanwhile, in order for a more complete analysis, we also translate the annotated English tweets into different languages (Spanish, Italian, and French) to generated training data for building sentiment analysis models for these languages. SenWave thus reveals the sentiment of global conversation in six different languages on COVID-19 (covering English, Spanish, French, Italian, Arabic and Chinese), followed the spread of the epidemic. The conversation showed a remarkably similar pattern of rapid rise and slow decline over time across all nations, as well as on special topics like the herd immunity strategies, to which the global conversation reacts strongly negatively. Overall, SenWave shows that optimistic and positive sentiments increased over time, foretelling a desire to seek, together, a reset for an improved COVID-19 world.

Understanding the Role of Affect Dimensions in Detecting Emotions from Tweets: A Multi-task Approach

“…Co-training especially helps in improving the performance of the classification task as we outperform the strongest baselines with 3.4%, 11%, and 3.9% gains in Jaccard Accuracy, Macro-F1, and Micro-F1 scores respectively on the AIT dataset [17]. We also achieve state-of-the-art results with 11.3% gains averaged over six different metrics on the SenWave dataset [27]. For the regression task, VADEC, when trained with SenWave, achieves 7.6% and 16.5% gains in Pearson Correlation scores over the current state-of-the-art on the EMOBANK dataset [5] for the Valence (V) and Dominance (D) affect dimensions respectively.…”

“…VADEC learns better shared representations by jointly training the two modules, that especially help in improving the performance of the classification task, thereby achieving state-of-the-art results on the AIT [17] and SenWave [27] datasets (Section 3.3). For the regression task, we achieve SOTA results on the EMOBANK dataset [5] for V and D dimensions (Section 3.4).…”

Semantic and sentiment analysis of selected Bhagavad Gita translations using BERT-based language framework

“…Although sentiment analysis can be done with model prediction that gives a positive/negative polarity score, this would not reveal much information about the nature of the sentiment as the positive/negative score can be vague due to different expressions used in translations. Hence, we use the hand-labelled SenWave dataset [112] which features 11 different sentiments labelled by a group of 50 experts for 10,000 tweets worldwide during COVID-19 pandemic in 2020.…”

“…The BERT base model is trained using the SenWave dataset [112] for a multi-label sentiment classification task. Although sentiment analysis can provide an idea regarding the differences in the translations, it does not provide information about the difference in the semantics (meaning) across the translations.…”

“…The pre-trained BERT-base uncased model is further tuned on the SenWave dataset [112] with training and test dataset created in a 90/10 % split on a dataset of 10,000 tweets using a batch size of 1 using 4 training epochs. We use the default hyper-parameters for the model and implement regularisation using a dropout [118] layer with dropout probability of 0.3 to avoid over-fitting.…”