Twitter User Representation using Weakly Supervised Graph Embedding

Travel behaviour market segmentations have become a popular method of identifying different types of car users, bicyclists or public transport users. However, while previous studies have looked at different types of users within single modes, such as the car, little research has explored the existence of traveller types transcending modes. The study presented here is an extension of an earlier segmentation study that distinguished travellers based on their individual preferences, yet did so independent of their current mode choice. The data came from a travel survey at a middle-sized UK university and were analysed using a combination of hierarchical and iterative partitioning methods. Crucially, however, the current study uses a different theoretical framework to previous segmentation research—goal framing theory—which may more adequately explain the findings than models used in the past such as the theory of planned behaviour. The findings supported earlier work, suggesting the presence of seemingly stable traveller types that cut across modes and can be distinguished based on gain, hedonic and normative goals. This has important implications for policies aimed at encouraging mode change which may have been too preoccupied with changing people’s attitudes rather than paying attention to people’s underlying travel preferences.

Social media platforms provide rich data sources in several domains. In mental health, individuals experiencing an Eating Disorder (ED) are often hesitant to seek help through conventional healthcare services. However, many people seek help with diet and body image issues on social media. To better distinguish at-risk users who may need help for an ED from those who are simply commenting on ED in social environments, highly sophisticated approaches are required. Assessment of ED risks in such a situation can be done in various ways, and each has its own strengths and weaknesses. Hence, there is a need for and potential benefit of a more complex multimodal approach. To this end, we collect historical tweets, user biographies, and online behaviours of relevant users from Twitter, and generate a reasonably large labelled benchmark dataset. Thereafter, we develop an advanced multimodal deep learning model called EDNet using these data to identify the different types of users with ED engagement (e.g., potential ED sufferers, healthcare professionals, or communicators) and distinguish them from those not experiencing EDs on Twitter. EDNet consists of five deep neural network layers. With the help of its embedding, representation and behaviour modeling layers, it effectively learns the multimodalities of social media. In our experiments, EDNet consistently outperforms all the baseline techniques by significant margins. It achieves an accuracy of up to 94.32% and F1 score of up to 93.91% F1 score. To the best of our knowledge, this is the first such study to propose a multimodal approach for user-level classification according to their engagement with ED content on social media.

Telehealth has played a critical role in supporting healthcare delivery during the COVID-19 pandemic. Mitigation practices of physical distancing and sheltering-in-place continue to impose a significant burden on face-to-face healthcare visits. As a result, telehealth adoption and utilization have increased significantly after the U.S. Department of Health and Human Services waived telehealth payment requirements. These changes have been rapid and affect both individuals and organizations, ranging from healthcare and technology to news and policy. We turn to Twitter, as a broad-reaching communication platform, to elucidate and characterize Telehealth communication patterns. While different aspects of tweeting about telehealth have been used to answer specific questions, we sought to understand the collective telehealth tweeting behavior as a system of different types of users tweeting with various intent and sentiment. Such an approach combines the typically independent analysis of structure, using social network analysis, and purpose or function of communication, with topic analysis and sentiment analysis. In this paper, we apply a systems-based approach-called hetero-functional graph theory-to comprehensively describe the structure and function of the Twitter Telehealth communication system. We modeled 13 different types of telehealth users and 15 combinations of tweet intent and sentiment, which together describe 195 system capabilities. We applied machine learning to classify individual user tweeting behavior for different classes of telehealth stakeholders. We identified several patterns of longitudinal tweeting behaviors that varied by system form and function. By applying an interdependent methodology to understand both form and function, we provide deeper insights into tweeting behaviors between different classes of stakeholders. In doing so, we achieve a greater understanding of the various needs of different stakeholders, which can be leveraged as a valuable resource in informing more equitable policies as telehealth becomes a more permanent and scaled healthcare delivery service.

Social media allows customers and prospects to communicate directly to your brand representative or about yourbrand with their friends. However, the obvious question is: who are the people interacting online and howengaged are they in online activities? This paper aims to answer this question based on a study regarding theonline activities of 236 social media users, by identifying different types of users, a segmentation of these usersand a linear model to examine how different predictors related to social networking sites have a positive impacton the respondents’ perception of online advertisements. The answer can help discover how to engage withdifferent types of audiences in order to maximize the effect of the online marketing strategy.

Considering the characteristics of different types of users in hybrid carsharing systems, in which sharing autonomous vehicles (SAVs) and conventional sharing cars (CSCs) coexist, a tailored pricing strategy (TPS) is proposed to maximize the operator’s profit and minimize all users’ costs. The fleet sizes and sizes of SAVs’ stations are also determined simultaneously. A bi-objective nonlinear programming model is established, and a genetic algorithm is applied to solve it. Based on the operational data in Lanzhou, China, carsharing users are clustered into three types. They are loyal users, losing users, and potential users, respectively. Results show the application of the TPS can help the operator increase profit and attract more users. The loyal users are assigned the highest price, while they still contribute the most to the operator’s profit with the highest number of carsharing trips. The losing users and potential users are comparable in terms of the number of trips, while the latter still makes more profit.

An information-pattern-based approach to novelty detection

Predicting the Adoption of Social Media: An Integrated Model and Empirical Study on Facebook Usage

Big data analytics in healthcare would be almost useless, without suitable tools allowing users see" them, and gain insight for their situated decisions. The VVH (Valuable Visualization in Healthcare) workshop focuses on the role of interactive data visualization tools by which people can make sense of healthcare data; these data include sensor data, the messages exchanged in social media, the emails between patients and their doctors, the content of patient records as well as the discussions among different specialists that led to such record content. All these data are used by different types of users, like doctors, nurses, policy makers and common citizens. The VVH workshop aims at contributing on: The assessment of the usability of advanced interactive tools of health-related data visualization; the assessment of the quality of the information and value for insight that these tools make available to their users; the collection of reports of either success stories or failures in the appropriation and use of complex and multidimensional healthcare datasets; the collection of methodological and design-oriented contributions that could share methods, techniques, and heuristics for the design of interactive tools and applications supporting data work, data telling and data interpretation in healthcare.

BackgroundA diagnosis of drug-resistant epilepsy is life changing for a family. Ketogenic diet therapy (KDT) can offer hope when other treatments have failed. However, it often requires a significant change in daily routine and dietary habits. This qualitative descriptive study aimed to explore families’ experiences of epilepsy and KDT. MethodsParents of a child aged ≤18 years with epilepsy, currently or recently treated with KDT, were recruited from the UK and internationally via UK Ketogenic Diet (KD) centres, charities, and social media. Semi-structured interviews were audio recorded, transcribed verbatim, anonymised, coded using Nvivo (V12), and inductive thematic analysis undertaken. ResultsTwenty-one parents participated. Four themes and 12 subthemes emerged: 1. ‘Epilepsy is all consuming’ explored the impact of epilepsy on the family. 2. ‘KD provides a window to new opportunities’ explores the motivators for KDT and positive outcomes. 3. ‘The reality of KD’ explores day to day life and how families adapt to KD. 4. ‘Looking to the future’ explores the factors that may make KD easier for families. All were glad their child trialled KD, even when less successful. The importance of a support network including family, friends, charity organisations and the KD team was evident across all themes. ConclusionsWe conclude with five recommendations to help support families in their management of KDT; Improved access to KDT and transition to adult services, access to quality education and support, enhanced variety of KD foods, regular social education and finally consideration of peer mentoring.

This study explores whether social roles mentioned in jihadist literature can be used linguistically to profile the behavior of its consumers. It seeks to empirically understand extremist forums, where such ideological communication is widely disseminated, and to identify different types of user behavior through their discussions in these communities. An Arabic-language corpus was constructed with 1050 users categorized into seven behavioral profiles derived from distinct patterns of communication observed among users in forum networks. Statistical key term extraction was used to find significant social roles referenced in users’ posts as these were proposed to differentiate their behavioral profiles. Multinomial logistic regression and post hoc tests were used to find the frequency of identified roles as the highest-ranking terms capable of prediction. Role terms produced high accuracy scores across classification experiments in identifying behavioral profiles (95% CI, 0.92–0.98), with varying intra- and inter-behavioral differentiation abilities emerging from the authority, religion, closeness, and conflict themes of social roles. This suggests simple and complex predictive potential exist in these constructs for separating profiles based on nuanced expressions of social roles by different types of users in extremist contexts, as was the case for distinguishing extremist organizations through their ideological communication in earlier work.

Towards optimal customized electricity pricing via iterative two-layer optimization for consumers and prosumers

Based on the investigation and analysis of typical users in Beijing and Guangzhou, a comparative analysis of the load characteristics of different types of users is carried out, and the basic characteristics of the load characteristics of various types of users are summarized. In the calculation example, the hourly load of the typical buildings in Beijing and Guangzhou is simulated and analyzed for the whole year. The load stability of the hotel and hospital buildings is poor and the continuity is strong. The load stability of shopping malls, residential and office buildings Stronger and weaker in persistence; the imbalance rate of the cumulative cooling and heating load in the summer hot winter and warm winter area and the cumulative cooling and heating load in the air conditioning season is higher, while the unbalance rate in the cold area is lower.

We propose a framework for describing, comparing and understanding tools for the mining of software repositories. The fundamental premise of this framework is that mining should be done by considering the specific needs of the users and the tasks to be supported by the mined information. First, different types of users have distinct needs, and these needs should be taken into account by tool designers. Second, the data sources available, and mined, will determine if those needs can be satisfied. Our framework is based upon three main principles: the type of user, the objective of the user, and the mined information. This framework has the following purposes: to help tool designers in the understanding and comparison of different tools, to assist users in the assessment of a potential tool; and to identify new research areas. We use this framework to describe several mining tools and to suggest future research directions.

Social media and blogging services have become extremely popular. Every day hundreds of millions of users share random thoughts, gossip, news, and thoughts on notable social issues. Users interact by following each other’s updates and passing along interesting pieces of information to their friends. Information therefore can diffuse widely and quickly through social links. Information propagation in networks such as Twitter and Facebook is unique, in that traditional media sources and word-of-mouth propagation coexist. The availability of digitally logged propagation events in social media helps one better understand how a wide range of factors that are essential in communication, such as user influence, tie strength, repeated exposures, mass media, and agenda setting, come into play in the way people generate and consume information in modern society. This chapter reviews the roles different types of users of social media play in information propagation as well as the resulting propagation patterns. It also discusses specific examples, including the spread of social conventions and identifying topic experts in social media, in an effort to bring about better understanding of the characteristics of propagation phenomena in large social networks.

Social media platform allow users to exercise their communication strategies for knowledge creation, sharing and dissemination, which can be use by government agencies as an opportunity to better serve their citizens. In this study, we examine the users' perspective of the Social Media User Behavior Model allowing investigating the social media-based knowledge creation, framing and targeting behaviors by different user types. 123 subjects responded to our Social Media Questionnaire, allowing us to measure perspectives of their social media usage (Preferences, Intensity and Functionality) as well as perspectives of their knowledge behaviors (creation, framing and targeting). The findings imply that users' perspectives support the Social Media User Behavior Model by supporting H1 and H3, connecting the knowledge behaviors targeting and creating to users' characteristics, especially social media Intensity and Functionality. The results means that decision makers can improve communication with the public by identifying social media intensive users and users that use social media for many (versus few) functions, and facilitate different information strategies management for different users types.