Language In Social Media Research Articles

An emotion and cognitive based analysis of mental health disorders from social media data

Mental disorders can severely affect quality of life, constitute a major predictive factor of suicide, and are usually underdiagnosed and undertreated. Early detection of signs of mental health problems is particularly important, since unattended, they can be life-threatening. This is why a deep understanding of the complex manifestations of mental disorder development is important. We present a study of mental disorders in social media, from different perspectives. We are interested in understanding whether monitoring language in social media could help with early detection of mental disorders, using computational methods. We developed deep learning models to learn linguistic markers of disorders, at different levels of the language (content, style, emotions), and further try to interpret the behavior of our models for a deeper understanding of mental disorder signs. We complement our prediction models with computational analyses grounded in theories from psychology related to cognitive styles and emotions, in order to understand to what extent it is possible to connect cognitive styles with the communication of emotions over time. The final goal is to distinguish between users diagnosed with a mental disorder and healthy users, in order to assist clinicians in diagnosing patients. We consider three different mental disorders, which we analyze separately and comparatively: depression, anorexia, and self-harm tendencies.

Developing and evaluating language‐based machine learning algorithms for inferring applicant personality in video interviews

AbstractOrganisations are increasingly relying on people analytics to aid human resources decision‐making. One application involves using machine learning to automatically infer applicant characteristics from employment interview responses. However, management research has provided scant validity evidence to guide organisations' decisions about whether and how best to implement these algorithmic approaches. To address this gap, we use closed vocabulary text mining on mock video interviews to train and test machine learning algorithms for predicting interviewee's self‐reported (automatic personality recognition) and interviewer‐rated personality traits (automatic personality perception). We use 10‐fold cross‐validation to test the algorithms' accuracy for predicting Big Five personality traits across both rating sources. The cross‐validated accuracy for predicting self‐reports was lower than large‐scale investigations using language in social media posts as predictors. The cross‐validated accuracy for predicting interviewer ratings of personality was more than double that found for predicting self‐reports. We discuss implications for future research and practice.

Auto-Off ID: Automatic Detection of Offensive Language in Social Media

As the popularity of social media grows, computer-mediated anonymity allows users to engage in activities that they would not do in real life. This makes users vulnerable to abuse through Internet platforms. Due to the enormous number of social media data, it is not possible to manually filter out the overflow of abusive content in online communities and social networking sites. The research work proposes a multi-level classification model that deploys various machine and deep learning models to effectively identify offensive content in a tweet. The proposed Auto-Off ID system is designed to build a system that classifies tweets as offensive or non-offensive; filters out and classifies offensive tweets as either targeted or non-targeted; filters out targeted tweets and identify mentions of individuals and organizations who have been bullied. The study is supported by the text analysis features with lexicon features using LIWC, POS tags for primary and secondary users, Twitter Tag Scores (TTS). This system is evaluated using a diverse choice of machine learning and deep learning models from which it is proved that C-LSTM outperform with an accuracy of 91.72% for offensive language identification; LDA + Logistic Regression training with SVM accuracy of 90.87% for offensive tweet classification.

Negation Detection on Mexican Spanish Tweets: The T-MexNeg Corpus

In this paper, we introduce the T-MexNeg corpus of Tweets written in Mexican Spanish. It consists of 13,704 Tweets, of which 4895 contain negation structures. We performed an analysis of negation statements embedded in the language employed on social media. This research paper aims to present the annotation guidelines along with a novel resource targeted at the negation detection task. The corpus was manually annotated with labels of negation cue, scope, and, event. We report the analysis of the inter-annotator agreement for all the components of the negation structure. This resource is freely available. Furthermore, we performed various experiments to automatically identify negation using the T-MexNeg corpus and the SFU ReviewSP-NEG for training a machine learning algorithm. By comparing two different methodologies, one based on a dictionary and the other based on the Conditional Random Fields algorithm, we found that the results of negation identification on Twitter are lower when the model is trained on the SFU ReviewSP-NEG Corpus. Therefore, this paper shows the importance of having resources built specifically to deal with social media language.

Social media language of healthcare super-utilizers

An understanding of healthcare super-utilizers’ online behaviors could better identify experiences to inform interventions. In this retrospective case-control study, we analyzed patients’ social media posts to better understand their day-to-day behaviors and emotions expressed online. Patients included those receiving care in an urban academic emergency department who consented to share access to their historical Facebook posts and electronic health records. Super-utilizers were defined as patients with more than six visits to the Emergency Department (ED) in a year. We compared posts by super-utilizers with a matched group using propensity scoring based on age, gender and Charlson comorbidity index. Super-utilizers were more likely to post about confusion and negativity (D = .65, 95% CI-[.38, .95]), self-reflection (D = .63 [.35, .91]), avoidance (D = .62 [.34, .90]), swearing (D = .52 [.24, .79]), sleep (D = .60 [.32, .88]), seeking help and attention (D = .61 [.33, .89]), psychosomatic symptoms, (D = .49 [.22, .77]), self-agency (D = .56 [.29, .85]), anger (D = .51, [.24, .79]), stress (D = .46, [.19, .73]), and lonely expressions (D = .44, [.17, .71]). Insights from this study can potentially supplement offline community care services with online social support interventions considering the high engagement of super-utilizers on social media.

Understanding Side Effects of Antidepressants: Large-scale Longitudinal Study on Social Media Data.

BackgroundAntidepressants are known to show heterogeneous effects across individuals and conditions, posing challenges to understanding their efficacy in mental health treatment. Social media platforms enable individuals to share their day-to-day concerns with others and thereby can function as unobtrusive, large-scale, and naturalistic data sources to study the longitudinal behavior of individuals taking antidepressants.ObjectiveWe aim to understand the side effects of antidepressants from naturalistic expressions of individuals on social media.MethodsOn a large-scale Twitter data set of individuals who self-reported using antidepressants, a quasi-experimental study using unsupervised language analysis was conducted to extract keywords that distinguish individuals who improved and who did not improve following the use of antidepressants. The net data set consists of over 8 million Twitter posts made by over 300,000 users in a 4-year period between January 1, 2014, and February 15, 2018.ResultsFive major side effects of antidepressants were studied: sleep, weight, eating, pain, and sexual issues. Social media language revealed keywords related to these side effects. In particular, antidepressants were found to show a spectrum of effects from decrease to increase in each of these side effects.ConclusionsThis work enhances the understanding of the side effects of antidepressants by identifying distinct linguistic markers in the longitudinal social media data of individuals showing the most and least improvement following the self-reported intake of antidepressants. One implication of this work concerns the potential of social media data as an effective means to support digital pharmacovigilance and digital therapeutics. These results can inform clinicians in tailoring their discussion and assessment of side effects and inform patients about what to potentially expect and what may or may not be within the realm of normal aftereffects of antidepressants.

Person-Centered Predictions of Psychological Constructs with Social Media Contextualized by Multimodal Sensing

Personalized predictions have shown promises in various disciplines but they are fundamentally constrained in their ability to generalize across individuals. These models are often trained on limited datasets which do not represent the fluidity of human functioning. In contrast, generalized models capture normative behaviors between individuals but lack precision in predicting individual outcomes. This paper aims to balance the tradeoff between one-for-each and one-for-all models by clustering individuals on mutable behaviors and conducting cluster-specific predictions of psychological constructs in a multimodal sensing dataset of 754 individuals. Specifically, we situate our modeling on social media that has exhibited capability in inferring psychosocial attributes. We hypothesize that complementing social media data with offline sensor data can help to personalize and improve predictions. We cluster individuals on physical behaviors captured via Bluetooth, wearables, and smartphone sensors. We build contextualized models predicting psychological constructs trained on each cluster's social media data and compare their performance against generalized models trained on all individuals' data. The comparison reveals no difference in predicting affect and a decline in predicting cognitive ability, but an improvement in predicting personality, anxiety, and sleep quality. We construe that our approach improves predicting psychological constructs sharing theoretical associations with physical behavior. We also find how social media language associates with offline behavioral contextualization. Our work bears implications in understanding the nuanced strengths and weaknesses of personalized predictions, and how the effectiveness may vary by multiple factors. This work reveals the importance of taking a critical stance on evaluating the effectiveness before investing efforts in personalization.

Depressive cognition on Twitter

Psychologists have long known that people with depression often have unhelpful, negative patterns of thinking, known as cognitive distortions. Bathina et al. now show that these thought patterns can be detected in the everyday language of social media and that individuals who report a diagnosis of depression express more cognitive distortions.

Text classification models for the automatic detection of nonmedical prescription medication use from social media

BackgroundPrescription medication (PM) misuse/abuse has emerged as a national crisis in the United States, and social media has been suggested as a potential resource for performing active monitoring. However, automating a social media-based monitoring system is challenging—requiring advanced natural language processing (NLP) and machine learning methods. In this paper, we describe the development and evaluation of automatic text classification models for detecting self-reports of PM abuse from Twitter.MethodsWe experimented with state-of-the-art bi-directional transformer-based language models, which utilize tweet-level representations that enable transfer learning (e.g., BERT, RoBERTa, XLNet, AlBERT, and DistilBERT), proposed fusion-based approaches, and compared the developed models with several traditional machine learning, including deep learning, approaches. Using a public dataset, we evaluated the performances of the classifiers on their abilities to classify the non-majority “abuse/misuse” class.ResultsOur proposed fusion-based model performs significantly better than the best traditional model (F1-score [95% CI]: 0.67 [0.64–0.69] vs. 0.45 [0.42–0.48]). We illustrate, via experimentation using varying training set sizes, that the transformer-based models are more stable and require less annotated data compared to the other models. The significant improvements achieved by our best-performing classification model over past approaches makes it suitable for automated continuous monitoring of nonmedical PM use from Twitter.ConclusionsBERT, BERT-like and fusion-based models outperform traditional machine learning and deep learning models, achieving substantial improvements over many years of past research on the topic of prescription medication misuse/abuse classification from social media, which had been shown to be a complex task due to the unique ways in which information about nonmedical use is presented. Several challenges associated with the lack of context and the nature of social media language need to be overcome to further improve BERT and BERT-like models. These experimental driven challenges are represented as potential future research directions.

Politeness in the Use of Language in Social Media

The use of social media in communication changes the way people express the idea represented through their language. This study aims to seek for the form of communication patterns in social media and to reveal causes of specific pattern language usage found in media, especially in office domain. The study was conducted using descriptive qualitative by describing the results of observations, interviews, and documentation. The data were collected by non-participatory observation method and note taking technique. Based on the results of the analysis of the data, it can be concluded that the communication patterns are realized in the form of: (1) various forms of language which do not meet the standard language, (2) the use of other non-verbal representations, such as emoticon, and (3) ecrononciation, that is, writing the spoken language. The choice of utterances is affected by relative status, social distance and range of imposition. As communicating to others, most utterances represent the obedience to politeness maxim, namely be friendly (sumanak), be considerate (tepa slira), suitable topic to event (empan papan), and pleasant voice true and good temper (nuju prana).

Detecting Offensive Language in Multi-Dialectal Arabic Social Media

Detecting Offensive Language in Multi-Dialectal Arabic Social Media

An Evaluation of the Accuracy of the Machine Translation Systems of Social Media Language

In this age of information technology, it has become possible for people all over the world to communicate in different languages through social media platforms with the help of machine translation (MT) systems. As far as the Arabic-English language pair is concerned, most studies have been conducted on evaluating the MT output for the standard varieties of Arabic, with fewer studies focusing on the vernacular or colloquial varieties. This study attempts to address this gap through presenting an evaluation of the performance of MT output for vernacular or colloquial Arabic in the social media domain. As it is currently the most widely used MT system, Google Translate (GT) has been chosen for evaluating the reliability of its output in the context of translating the Arabic colloquial language (i.e., Egyptian/Cairene Arabic variety) used in social media into English. With this goal in mind, a corpus consisting of Egyptian dialectal Arabic sentences were collected from social media networks, i.e., Facebook and Twitter, and then fed into GT system. The GT output was then evaluated by three human translators to assess their accuracy of translation in terms of adequacy and fluency. The results of the study show that several translation problems have been spotted for GT output. These problems are mainly concerned with wrong equivalents, inappropriate additions and deletions, and transliteration for out-of-vocabulary (OOV) words, which are mostly due to the literal translation of the Arabic vernacular sentences into English. This can be due to the fact that Arabic vernacular varieties are different from the standard language for which MT systems have been basically developed. This, consequently, necessitates the need to upgrade such MT systems to deal with the vernacular varieties.

Is neural always better? SMT versus NMT for Dutch text normalization

Social media texts have become one of the most used forms of written language and a valuable source of information for companies. However, despite the wealth of information hidden in social media language for profiling, information extraction and sentiment analysis purposes, this user-generated content is also characterized by non-standard language. This makes it difficult to process using standard NLP techniques as these are typically trained on standard text. Text normalization is often used as a preprocessing step to overcome this problem. In this work, we take a machine translation perspective on text normalization and investigate both statistical and neural methods. We perform text normalization in a low-resource scenario and report both SMT and NMT experiments for three different Dutch user-generated text types: tweets, message board posts and text messages. Furthermore, we look into overcoming the over-normalization problem in NMT using different techniques, viz. teacher forcing, copy mechanism and byte pair encoding. Our results reveal that even though the SMT approach obtains the best results, the NMT system using CopyNet shows promising results even in this low-resource setting, solving more normalization issues than SMT although not solving all the over-normalization problems.

Pemanfaatan Bahasa Iklan Sebagai Penunjang Media promosi di Jalan Akasia

Warmadewa University has had a significant transformation to its surrounding which used to be agricultural land, currently this land is occupied by several traders who offer their products and services especially on Akasia street. There are differences when viewing from an economic point of view before the pandemic which has been going on for the last few months which has caused a drastic decline in turnover of traders around the Warmadewa University. Based on the visions and missions of Warmadewa University to increase the quantity and quality of the community service that is the prosperity of the traders. Then, the solutions that can be offered is to apply the method, namely the socialization of advertising languages to attract consumers through print and online media with the hope of a revival of the economy for traders in the midst of this pandemic. The results of this study prove that Community Service Activities with the target of the traders at Akasia Street have gotten better by increasing partners' knowledge in promotional strategy by applying adveritising language through print media such as distributing printed and online media such as distribution of product/service advertising banners in the form of education in the use of advertising language in social media accounts such as WhatsApp, Instagram and Facebook.

Psychosocial Effects of the COVID-19 Pandemic: Large-scale Quasi-Experimental Study on Social Media.

BackgroundThe COVID-19 pandemic has caused several disruptions in personal and collective lives worldwide. The uncertainties surrounding the pandemic have also led to multifaceted mental health concerns, which can be exacerbated with precautionary measures such as social distancing and self-quarantining, as well as societal impacts such as economic downturn and job loss. Despite noting this as a “mental health tsunami”, the psychological effects of the COVID-19 crisis remain unexplored at scale. Consequently, public health stakeholders are currently limited in identifying ways to provide timely and tailored support during these circumstances.ObjectiveOur study aims to provide insights regarding people’s psychosocial concerns during the COVID-19 pandemic by leveraging social media data. We aim to study the temporal and linguistic changes in symptomatic mental health and support expressions in the pandemic context.MethodsWe obtained about 60 million Twitter streaming posts originating from the United States from March 24 to May 24, 2020, and compared these with about 40 million posts from a comparable period in 2019 to attribute the effect of COVID-19 on people’s social media self-disclosure. Using these data sets, we studied people’s self-disclosure on social media in terms of symptomatic mental health concerns and expressions of support. We employed transfer learning classifiers that identified the social media language indicative of mental health outcomes (anxiety, depression, stress, and suicidal ideation) and support (emotional and informational support). We then examined the changes in psychosocial expressions over time and language, comparing the 2020 and 2019 data sets.ResultsWe found that all of the examined psychosocial expressions have significantly increased during the COVID-19 crisis—mental health symptomatic expressions have increased by about 14%, and support expressions have increased by about 5%, both thematically related to COVID-19. We also observed a steady decline and eventual plateauing in these expressions during the COVID-19 pandemic, which may have been due to habituation or due to supportive policy measures enacted during this period. Our language analyses highlighted that people express concerns that are specific to and contextually related to the COVID-19 crisis.ConclusionsWe studied the psychosocial effects of the COVID-19 crisis by using social media data from 2020, finding that people’s mental health symptomatic and support expressions significantly increased during the COVID-19 period as compared to similar data from 2019. However, this effect gradually lessened over time, suggesting that people adapted to the circumstances and their “new normal.” Our linguistic analyses revealed that people expressed mental health concerns regarding personal and professional challenges, health care and precautionary measures, and pandemic-related awareness. This study shows the potential to provide insights to mental health care and stakeholders and policy makers in planning and implementing measures to mitigate mental health risks amid the health crisis.

Fungsi bahasa dan makna pragmatik dalam media sosial untuk memerangi Covid-19: Kajian etnopragmatik

This article discusses the function of language and pragmatic meaning in social media to combat Covid-19. Theoretical references used are ethnopragmatics. Ethnographic studies basically examine language from a cultural aspect. Cultural aspects include knowledge, behavior, and artifacts. Meanwhile, pragmatics is a study of language use on the basis of context. This research uses data sources from Instagram, Facebook, WA, and email. Research data in the form of language use in social media above. Data from Instagram and Facebook are downloaded from the internet, while data from email is collected through the direction of researchers with guidance on the instrument. Data collection techniques are done through reading and taking notes. Concrete data analysis techniques are carried out through data identification, data classification, data interpretation, and reporting in the form of journal articles. The findings of research on language subfunctions include asking, inviting or influencing, being sarcastic, motivating and criticizing, and saying thank you. Meanwhile, the pragmatic meaning used in social media includes indirectly advising, persuading, indirectly reminding, subtly insinuating, inviting, complaining, and give advice.

MISMIS: Misinformation and Miscommunication in social media: aggregating information and analysing language

The general objectives of the project are to address and monitor misinformation (biased and fake news) and miscommunication (aggressive language and hate speech) in social media, as well as to establish a high quality methodological standard for the whole research community (i) by developing rich annotated datasets, a data repository and online evaluation services; (ii) by proposing suitable evaluation metrics; and (iii) by organizing evaluation campaigns to foster research on the above issues.

“#JeSuisSirCornflakes”: Racialization and resemiotization in French nationalist Twitter

Abstract After the 2016 spelling reforms deleted the accent circumflex from some French vowels, on right-wing French Twitter, the circonflexe reappeared in the center of the French flag – echoing the flag of Vichy France. Tweets with the hashtag #JeSuisCirconflexe resemiotized the accent circumflex as icon of a lost Frenchness, or voiced the racial other in a colonial faux pidgin to frame them as illiterate and brutish. Drawing on research on resemiotization (Leppänen, Sirpa, Samu Kytölä, Henna Jousmäki, Saija Peuronen & Elina Westinen. 2014. Entextualization and resemiotization as resources for identification in social media. In The language of social media, 112–136. London: Palgrave Macmillan) and raciolinguistics (Flores, Nelson & Jonathan Rosa. 2015. Undoing appropriateness: Raciolinguistic ideologies and language diversity in education. Harvard Educational Review, 85(2). 149–171), this article explores how constructions of mock youth French use raciolinguistic tropes to imagine a language of social decline, connecting linguistic purism to racist myths of white genocide and the great replacement. Despite this, youth invert the imagination of their illiteracy, using playful language and satirizing white speech (Rosa, Jonathan. 2016b. Standardization, racialization, languagelessness: Raciolinguistic ideologies across communicative contexts. Journal of Linguistic Anthropology 26(2). 162–183) to contest French nationalism – reframing #JeSuisCirconflexe as #JeSuisSirCornflakes.

Naturally occurring language as a source of evidence in suicide prevention

We discuss computational language analysis as it pertains to suicide prevention research, with an emphasis on providing non-technologists with an understanding of key issues and, equally important, considering its relation to the broader enterprise of suicide prevention. Our emphasis here is on naturally occurring language in social media, motivated by its non-intrusive ability to yield high-value information that in the past has been largely unavailable to clinicians.

Tracking Fluctuations in Psychological States using Social Media Language: A Case Study of Weekly Emotion

Personality psychologists are increasingly documenting dynamic, within–person processes. Big data methodologies can augment this endeavour by allowing for the collection of naturalistic and personality–relevant digital traces from online environments. Whereas big data methods have primarily been used to catalogue static personality dimensions, here we present a case study in how they can be used to track dynamic fluctuations in psychological states. We apply a text–based, machine learning prediction model to Facebook status updates to compute weekly trajectories of emotional valence and arousal. We train this model on 2895 human–annotated Facebook statuses and apply the resulting model to 303 575 Facebook statuses posted by 640 US Facebook users who had previously self–reported their Big Five traits, yielding an average of 28 weekly estimates per user. We examine the correlations between model–predicted emotion and self–reported personality, providing a test of the robustness of these links when using weekly aggregated data, rather than momentary data as in prior work. We further present dynamic visualizations of weekly valence and arousal for every user, while making the final data set of 17 937 weeks openly available. We discuss the strengths and drawbacks of this method in the context of personality psychology's evolution into a dynamic science. © 2020 European Association of Personality Psychology