Overview of the FIRE 2022 track: Information Retrieval from Microblogs during Disasters (IRMiDis)

Microblogging sites such as Twitter play an important role in dealing with various mass emergencies including natural disasters and pandemics. The FIRE 2022 track on Information Retrieval from Microblogs during Disasters (IRMiDis) focused on two important tasks – (i) to detect the vaccine-related stance of tweets related to COVID-19 vaccines, and (ii) to detect reporting of COVID-19 symptom in tweets. © 2022 Copyright for this paper by its authors.

Multiliteracies in the Analysis of a Multimodal Tweet from a Former Brazilian Minister of Education

The COVID-19 pandemic abruptly led millions of teachers and students in Brazil to migrate massively, quickly, and at scale to online remote teaching. This created a strong tension between different sectors of society and rekindled (old) beliefs, values, and prejudices related to the use of new technologies in education. On the one hand are vehement defenders for adoption of these technologies at schools;on the other are those who reject them, as they consider that transitioning from in-presence to online teaching at scale is a very difficult and highly complex undertaking for education systems. In this chapter, one seeks to discuss how the perspective of multiliteracies, updated for the currently pervasively digital age, can contribute to understanding the clash between defense and resistance to new technologies at schools. To do so, first, this article will explore the main features and concepts of the theory of multiliteracies. Second, in order to highlight the close relationship between multiliteracies and education, the article analyzes an example of a multimodal tweet posted on Twitter by a former Minister of Education in Brazil, addressing the Brazilian public school setting of online remote teaching.

Classification of Covid-19 Vaccine Opinion and Detection of Symptom-Reporting on Twitter Using Neural Networks

This paper describes my work for the Information Retrieval from Microblogs during Disasters.This track is divided into two sub-tasks. Task 1 is to build an effective classifier for 3-class classification on tweets with respect to the stance reflected towards COVID-19 vaccines.Task 2 is to devise an effective classifier for 4-class classification on tweets that can detect tweets that report someone experiencing COVID-19 symptoms.This paper proposes a classification method based on MLP classifier model.The evaluation shows the performance of our approach, which achieved 0.304 on F-Score in Task 1 and 0.239 on F-Score in Task 2. © 2022 Copyright for this paper by its authors.

Forecasting COVID-19 Vaccination Rates using Social Media Data

The COVID-19 pandemic has had a profound impact on the global community, and vaccination has been recognized as a crucial intervention. To gain insight into public perceptions of COVID-19 vaccines, survey studies and the analysis of social media platforms have been conducted. However, existing methods lack consideration of individual vaccination intentions or status and the relationship between public perceptions and actual vaccine uptake. To address these limitations, this study proposes a text classification approach to identify tweets indicating a user's intent or status on vaccination. A comparative analysis between the proportions of tweets from different categories and real-world vaccination data reveals notable alignment, suggesting that tweets may serve as a precursor to actual vaccination status. Further, regression analysis and time series forecasting were performed to explore the potential of tweet data, demonstrating the significance of incorporating tweet data in predicting future vaccination status. Finally, clustering was applied to the tweet sets with positive and negative labels to gain insights into underlying focuses of each stance. © 2023 ACM.

Need for Vision: A data-centric approach towards analysing impact of COVID-19

From the beginning of 2020, we saw a rise of a new virus called the Coronavirus and ultimately a pandemic that anyone reading this paper must have been through. With the rise of COVID,many vaccines were found, the global vaccination drive as a result of this naturally fueled a possibility of Pro-Vaxxers and Anti-Vaxxers strongly expressing their support and concerns regarding the vaccines on social media platforms and along with this came up the need of quick identification of people who are experiencing COVID-19 symptoms. So in this paper, an effort has been made to facilitate the understanding of all these complications and help the concerned authorities. With the help of data in the form of Covid-19 tweets, a (machine-learning) classifier has been built which can classify users as per their vaccine related stance and also classify users who have reported their symptoms through tweets. © FIRE 2022: Forum for Information Retrieval Evaluation.

CTC: COVID-19 Tweet Classification using CT-BERT

CTC is my submitted work to the Information Retrieval from Microblogs during Disasters (IRMiDis) Track at the Forum for Information Retrieval Evaluation (FIRE) 2022. Coronavirus disease (COVID-19) is an infectious disease caused by the SARS-CoV-2 virus. Most people infected with the virus experience a mild to moderate respiratory illness and recover without requiring special treatment. However, some become seriously ill and require medical attention. Vaccines against coronavirus and prompt reporting of symptoms saved many lives during the pandemic. The analysis of COVID-19-related tweets can provide valuable insights regarding the stance of people toward the new vaccine. It can also help the authorities to plan their strategies based on people's opinions about the vaccine and ensure the effectiveness of vaccination campaigns. Tweets describing symptoms can also aid in identifying high-alert zones and determining quarantine regulations. The IRMiDis track focuses on these COVID-19-related tweets that flooded Twitter. I developed an effective classifier for both Tasks 1 and 2. The evaluation score of my submitted run is reported in terms of accuracy and macro-F1 score. I achieved an accuracy of 0.770, a macro-F1 score of 0.773 in Task 1, and an accuracy of 0.820, a macro-F1 score of 0.746 in Task 2. I enjoyed the first rank among other submissions in both the tasks. © 2022 Copyright for this paper by its authors.

Regret, Delete, (Do Not) Repeat: An Analysis of Self-Cleaning Practices on Twitter After the Outbreak of the COVID-19 Pandemic

During the outbreak of the COVID-19 pandemic, many people shared their symptoms across Online Social Networks (OSNs) like Twitter, hoping for others' advice or moral support. Prior studies have shown that those who disclose health-related information across OSNs often tend to regret it and delete their publications afterwards. Hence, deleted posts containing sensitive data can be seen as manifestations of online regrets. In this work, we present an analysis of deleted content on Twitter during the outbreak of the COVID-19 pandemic. For this, we collected more than 3.67 million tweets describing COVID-19 symptoms (e.g., fever, cough, and fatigue) posted between January and April 2020. We observed that around 24% of the tweets containing personal pronouns were deleted either by their authors or by the platform after one year. As a practical application of the resulting dataset, we explored its suitability for the automatic classification of regrettable content on Twitter. © 2023 Owner/Author.

Improving Sentiment Prediction of Textual Tweets Using Feature Fusion and Deep Machine Ensemble Model

Widespread fear and panic has emerged about COVID-19 on social media platforms which are often supported by falsified and altered content. This mass hysteria creates public anxiety due to misinformation, misunderstandings, and ignorance of the impact of COVID-19. To assist health professionals in addressing this epidemic more appropriately at the onset, sentiment analysis can potentially help the authorities for devising appropriate strategies. This study analyzes tweets related to COVID-19 using a machine learning approach and offers a high-accuracy solution. Experiments are performed involving different machine and deep learning models along with various features such as Word2vec, term-frequency, term-frequency document frequency, and feature fusion of both feature-generating approaches. The proposed approach combines the extra tree classifier and convolutional neural network and uses feature fusion to achieve the highest accuracy score of 99%. The proposed approach obtains far better results than existing sentiment analysis approaches. © 2023 by the authors.

Scanxiety Conversations on Twitter: Observational Study.

BACKGROUND: Scan-associated anxiety (or "scanxiety") is commonly experienced by people having cancer-related scans. Social media platforms such as Twitter provide a novel source of data for observational research. OBJECTIVE: We aimed to identify posts on Twitter (or "tweets") related to scanxiety, describe the volume and content of these tweets, and describe the demographics of users posting about scanxiety. METHODS: We manually searched for "scanxiety" and associated keywords in cancer-related, publicly available, English-language tweets posted between January 2018 and December 2020. We defined "conversations" as a primary tweet (the first tweet about scanxiety) and subsequent tweets (interactions stemming from the primary tweet). User demographics and the volume of primary tweets were assessed. Conversations underwent inductive thematic and content analysis. RESULTS: A total of 2031 unique Twitter users initiated a conversation about scanxiety from cancer-related scans. Most were patients (n=1306, 64%), female (n=1343, 66%), from North America (n=1130, 56%), and had breast cancer (449/1306, 34%). There were 3623 Twitter conversations, with a mean of 101 per month (range 40-180). Five themes were identified. The first theme was experiences of scanxiety, identified in 60% (2184/3623) of primary tweets, which captured the personal account of scanxiety by patients or their support person. Scanxiety was often described with negative adjectives or similes, despite being experienced differently by users. Scanxiety had psychological, physical, and functional impacts. Contributing factors to scanxiety included the presence and duration of uncertainty, which was exacerbated during the COVID-19 pandemic. The second theme (643/3623, 18%) was the acknowledgment of scanxiety, where users summarized or labeled an experience as scanxiety without providing emotive clarification, and advocacy of scanxiety, where users raised awareness of scanxiety without describing personal experiences. The third theme was messages of support (427/3623, 12%), where users expressed well wishes and encouraged positivity for people experiencing scanxiety. The fourth theme was strategies to reduce scanxiety (319/3623, 9%), which included general and specific strategies for patients and strategies that required improvements in clinical practice by clinicians or health care systems. The final theme was research about scanxiety (50/3623, 1%), which included tweets about the epidemiology, impact, and contributing factors of scanxiety as well as novel strategies to reduce scanxiety. CONCLUSIONS: Scanxiety was often a negative experience described by patients having cancer-related scans. Social media platforms like Twitter enable individuals to share their experiences and offer support while providing researchers with unique data to improve their understanding of a problem. Acknowledging scanxiety as a term and increasing awareness of scanxiety is an important first step in reducing scanxiety. Research is needed to guide evidence-based approaches to reduce scanxiety, though some low-cost, low-resource practical strategies identified in this study could be rapidly introduced into clinical care.

Rumor Detection on Social Media Using Deep Learning Algorithms with Fuzzy Inference System for Healthcare Analytics System Using Covid-19 Dataset

Spreading rumors on social media is a phenomenon that has destructive implication of societal interaction, diverts attention toward destructive behavior. The impact will be more influenced in healthcare management. This research aims to detect the rumors and identify the sources using deep learning algorithms. In our proposed system, after pre-processing, the tweet comments are extracted from topics and ranked as deny, support, query and comment. Then the comments are classified as positive, negative and neutral using Artificial Neural Network Neuro-fuzzy Inference System Spline-based pi-shaped Membership Function (ANISPIMF). Then the negative comments are classified into offensive, violence, misogyny and hate mongering by using Improved Deep Learning Neural Network (IDLNN) which is the combination of Deep Neural Network with Cuckoo Search–Flower Pollination Algorithm to optimize the weight values. The optimized ANISPIMF performs very well for the COVID-19 dataset in terms of Accuracy, Precision and Recall. The proposed system attains better performance and efficiency when weighted against prevailing methodologies — regarding the performance measures, there is an improvement of accuracy by 0.6%, recall by 0.7%, and precision by 1%, together with an F1-score of 1.2% than the Multiloss Hierarchical Bi-LSTM with Attenuation Factor (MHA). [ABSTRACT FROM AUTHOR] Copyright of International Journal of Computational Intelligence & Applications is the property of World Scientific Publishing Company and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

A Hybrid CNN-LSTM Model with Word-Emoji Embedding for Improving the Twitter Sentiment Analysis on Indonesia's PPKM Policy

The policy of limiting community mobilization is implemented to reduce the daily rate of COVID-19. However, a high-accuracy sentiment analysis model can determine public sentiment toward such policies. Our research aims to improve the accuracy of the LSTM model on sentiment analysis of the Jakarta community towards PPKM using Indonesian language Tweets with emoji embedding. The first stage is modeling using the hybrid CNN-LSTM model. It is a combination between CNN and LSTM. The CNN model cites word embedding and emoji embedding features that reflect the dependence on temporary short-term sentiment. At the same time, LSTM builds long-term sentiment relationships between words and emojis. Next, the model evaluation uses Accuracy, Loss, the receiver operating curve (ROC), the precision and recall curve, and the area under curve (AUC) value to see the performance of the designed model. Based on the results of the tests, we conclude that the CNN-LSTM Hybrid Model performs better with the words+emoji dataset. The ROC AUC is 0.966, while the precision-recall curve AUC is 0.957. © 2022 IEEE.

Twitter Data Sentiment Analysis of COVID-19 Vaccination using Machine Learning

The expansion of the web is accelerating, which helps encourage the creation of fresh ideas. In today's internet era, we must suggest techniques to filter out various information. Social media sentiment analysis based on Twitter data can monitor the real-Time monitoring of the COVID-19 vaccine. In this way, relevant organizations or governments can take proactive steps to address misinformation and inappropriate behaviour around the COVID-19 vaccine, which threatens the success of the national vaccination campaign. The purpose of this research is to determine if there is a link between how people feel about the COVID-19 vaccine on Twitter and how many people actually get vaccinated against it. This study uses the COVID-19 All Vaccines Tweet dataset sourced from Kaggle. This research Identifies public sentiment, emotion, word usage, and trend of all filtered tweets. The results show that there are 31% positive tweets, 10% negative tweets, and 58% neutral tweets. Tweets with neutral subjective valence tend to cluster in the middle of the polarity scale (between-1 and +1), whereas tweets with strong subjective valence are spread across the scale. © 2022 IEEE.

Clothing consumption during the COVID-19 pandemic: Evidence from mining Tweets

Understanding how consumers have shifted in clothing consumption in the midst of the global COVID-19 pandemic is critical for fashion clothing brands and businesses to identify what value means to consumers to locate growth opportunities. This exploratory study intends to provide a picture of consumers' clothing consumption evolution while going through the pandemic crisis. We take a viewpoint that integrates the perspectives of life status changes and stress coping to examine consumers' responses to clothing consumption during the COVID-19 global pandemic. A total of 68,511 relevant tweets were collected from January 1, 2020, through September 31, 2020. Sentiment and content analysis identified five themes which are revealed by 16 topics associated with clothing consumption over the phases of pre-lockdown, lockdown, and reopening. Pent-up demand for clothing products and changed clothing consumption habits were identified. Our findings provide evidence that consumption change is the fundamental mechanism of stress coping. (PsycInfo Database Record (c) 2022 APA, all rights reserved)

Which papers cited which tweets? An exploratory analysis based on Scopus data

Many altmetric studies have analyzed which papers were mentioned how often on Twitter (one of the most important altmetrics sources). In order to study the potential relevance of tweets from another perspective, we investigate which tweets were cited in papers. If many tweets were cited in publications, this might demonstrate that tweets have substantial and useful content. Overall, a rather low number of citations to tweets (n=13,149) by less than 7,000 papers was found. Most tweets do not seem to be cited because of any cognitive influence they might have had on studies;they rather were study objects. Thus, this study does not support a high relevance of tweets (for research). Most of the papers that cited tweets are from the subject areas Social Sciences, Arts and Humanities, and Medicine. Most of the papers cited only one tweet. Up to 65 tweets cited in a single paper were found. An author keyword analysis revealed that the single largest topic seems to be the COVID-19/corona pandemic. © 2023 Elsevier Ltd

Memory Augmented Distributed Monte Carlo Tree Search Algorithm-Based Content Popularity Aware Content Recommendation Using Content Centric Networks

The rapid advancements of social media networks have created the problem of overloaded information. As a result, the service providers push multiple redundant contents and advertisements to the users without adequate analysis of the user interests. The content recommendation without user interests reduces the probability of users reading them and the wastage rate of network load increases. This problem can be alleviated by providing accurate content recommendations with consideration of users' precise interests and content similarity. Content centric networking has been developed as the trending framework to satisfy these requirements and improve access to relevant information and reception by the desired user. The uses of message entity by giving a proper name, the users' real-time interests are identified and then the accurate and popular contents with high contextual similarity are recommended. An efficient content recommendation scheme is presented in this paper using Memory Augmented Distributed Monte Carlo Tree Search (MAD-MCTS) algorithm for ensuring minimum energy consumption in the CCN. The big data context of the users' social media data is considered in this study so that the complexity can be visualized and controlled to minimize the network complexities. Experiments are conducted on a benchmark as well as an offline collected Twitter dataset on Covid-19 and the results implied that the accuracy and convergence of the proposed MAD-MCTS outperform the other content recommendation algorithms. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Geo-sentiment trends analysis of tweets in context of economy and employment during COVID-19.

To effectively design policies and implement measures for addressing problems faced by people during these difficult times of pandemic, it is critical to have a clear vision of the problems people are freely talking about. One of the ways is to analyze social media feeds e.g., tweets, which has become one of the primary ways people express their views on various socioeconomic issues and on-ground effectiveness of measures adopted to address these issues. In this work, we attempt to uncover various socioeconomic issues, which are giving rise to negative and positive sentiments and their trends across geographies over a course of one year of the pandemic. We also try identifying similarities and differences in opinions as they vary across gender as the time passes through the crisis. Many previous works have analyzed sentiments in context of vaccines, fatalities, and lockdowns; however, socioeconomic issues did not receive full attention. We found that sentiments of people with respect to economy are negative across geographies during starting of pandemic. Thereafter, gradually sentiments lift towards positive direction reflecting a sense of improvement in situation. Females appeared to have slightly different concerns and hopes in comparison to males and especially across globe people expressed positive sentiments during new year time. Finally, this work, together with many other similar works on social media analysis gives ground for wide scale adoption of geo-temporal sentiments trend analysis of social media as a tool for uncovering key concerns and effectiveness of measures.

SCLAVOEM: hyper parameter optimization approach to predictive modelling of COVID-19 infodemic tweets using smote and classifier vote ensemble.

Fake COVID-19 tweets are dangerous since they are misinformative, completely inaccurate, as threatening the efforts for flattening the pandemic curve. Thus, aside the COVID-19 pandemic, dealing with fake news and myths about the virus constitute an infodemic issue, which must be tackled by ensuring only valid information. In this context, this study proposed the Synthetic Minority Over-Sampling Technique (SMOTE) and the classifier vote ensemble (SCLAVOEM) method as a fake news classifier and a hyper parameter optimization approach for predictive modelling of COVID-19 infodemic tweets. Hyper parameter optimization variables were deployed across specific points of the proposed model and a minority oversampling of training sets was applied within imbalanced class representations. Experimental applications by the SCLAVOEM for COVID-19 infodemic prediction returned 0.999 and 1.000 weighted averages for F-measure and area under curve (AUC), respectively. Thanks to the SMOTE, the performance increases of 3.74 and 1.11%; 5.05 and 0.29%; 4.59 and 8.05% was seen in three different data sets. Eventually, the SCLAVOEM provided a framework for predictive detecting 'fake tweets' and three classifiers: 'positive', 'negative' and 'click-trap' (piège à clics). It is thought that the model will automatically flag fake information on Twitter, hence protecting the public from inaccurate and information overload.

Identifying informative tweets during a pandemic via a topic-aware neural language model.

Every epidemic affects the real lives of many people around the world and leads to terrible consequences. Recently, many tweets about the COVID-19 pandemic have been shared publicly on social media platforms. The analysis of these tweets is helpful for emergency response organizations to prioritize their tasks and make better decisions. However, most of these tweets are non-informative, which is a challenge for establishing an automated system to detect useful information in social media. Furthermore, existing methods ignore unlabeled data and topic background knowledge, which can provide additional semantic information. In this paper, we propose a novel Topic-Aware BERT (TABERT) model to solve the above challenges. TABERT first leverages a topic model to extract the latent topics of tweets. Secondly, a flexible framework is used to combine topic information with the output of BERT. Finally, we adopt adversarial training to achieve semi-supervised learning, and a large amount of unlabeled data can be used to improve inner representations of the model. Experimental results on the dataset of COVID-19 English tweets show that our model outperforms classic and state-of-the-art baselines.

Predicting COVID-19 Related Tweets Using Ensemble of Transformers Models

In the past few years, COVID-19 has been consid-ered one of the most dangerous pandemics in several countries. There is a lot of information circulating on social media platforms about COVID-19, some of it is reliable, while others may be exag-gerated or unfounded. Using machine learning-driven sentiment analysis is considered a valuable tool that helps understand the community's feelings regarding many issues like the COVID-19 outbreak. Developing an accurate model that can assess if a tweet is about COVID-19 is a challenging task. This study aims to classify the tweets whether it is about COVID-19 or not using deep learning and transformers models. The developed model improves the gathering of tweets data about the COVID-19 epidemic without relying only on keywords such as 'covid' or 'coronavirus'. In this work, we proposed the best model based on an ensemble method that effectively combines three models which are: BERTweet-covid19-base-cased, BERTweet, and RoBERTa. We applied the models to the data set provided by the Zindi community. The best results were achieved over the tested dataset in terms of Log-Loss with a minimum value of 0.154,0.174,0.170, and 0.191 for the proposed ensemble model, BERTweet-covid19-base-cased, BERTweet, and RoBERTa respectively. Our proposed model is ranked first among all the participant teams. © 2022 IEEE.

Overview of the FIRE 2022 track: Information Retrieval from Microblogs during Disasters (IRMiDis)

Microblogging sites such as Twitter play an important role in dealing with various mass emergencies including natural disasters and pandemics. Over the last several years, the track on Information Retrieval from Microblogs during Disasters (IRMiDis), organized as part of the FIRE conference series, has provided annotated datasets for developing ML/NLP techniques for utilizing microblogs for various practical tasks that would help authorities better deal with disaster situations. In particular, the FIRE 2022 IRMiDis track focused on two important tasks-(i) to detect the vaccine-related stance of tweets related to COVID-19 vaccines, and (ii) to detect reporting of COVID-19 symptom in tweets. © 2022 Owner/Author.