Tweet Messages Research Articles

A deep-learning model for urban traffic flow prediction with traffic events mined from twitter

Short-term traffic parameter forecasting is critical to modern urban traffic management and control systems. Predictive accuracy in data-driven traffic models is reduced when exposed to non-recurring or non-routine traffic events, such as accidents, road closures, and extreme weather conditions. The analytical mining of data from social networks – specifically twitter – can improve urban traffic parameter prediction by complementing traffic data with data representing events capable of disrupting regular traffic patterns reported in social media posts. This paper proposes a deep learning urban traffic prediction model that combines information extracted from tweet messages with traffic and weather information. The predictive model adopts a deep Bi-directional Long Short-Term Memory (LSTM) stacked autoencoder (SAE) architecture for multi-step traffic flow prediction trained using tweets, traffic and weather datasets. The model is evaluated on an urban road network in Greater Manchester, United Kingdom. The findings from extensive empirical analysis using real-world data demonstrate the effectiveness of the approach in improving prediction accuracy when compared to other classical/statistical and machine learning (ML) state-of-the-art models. The improvement in predictive accuracy can lead to reduced frustration for road users, cost savings for businesses, and less harm to the environment.

Internet of things-based fog and cloud computing technology for smart traffic monitoring

Internet of Things (IoT) is changing the world by connecting billions of physical and virtual objects with distinctive identities to the Internet. This fusion results in generating huge volumes of data that might not be manageable using today's storage and data analytics technologies. Although cloud computing offers services to tackle this issue at infrastructural level, its efficiency for time sensitive applications (e.g. oil, gas, and traffic monitoring) is still questionable. Arguably, transferring massive amount of data to the cloud for storage and processing may lead to cloud overloading and saturation of network bandwidth. In this study, an integrated fog and cloud computing framework is introduced to overcome the limitations of real-time analytics, latency and network congestion of basic cloud services for traffic monitoring. The proposed approach is implemented to prototype a smart traffic monitoring system (STMS). The proposed monitoring system is designed for congestion monitoring and traffic light management. It can also be tuned to detect traffic incidents that requires immediate assistance during congestion. In this framework, a tiny computer-on-module serves as a fog node to collect real-time data from geographically distributed sensors and to transfer it to the cloud for storage and processing. The results show the efficiency of the fog network in improving the performance of the cloud platform in terms of reducing the response time and increasing the bandwidth. Furthermore, the proposed integrated fog and cloud framework is interfaced with Tweeter to send alerts about traffic congestion to be subscribed users in the form of Tweet messages .

Emotions in Social Media: An Analysis of Tweet Responses to MH370 Search Suspension Announcement

Considered one of the deadliest incidents in the history of aviation crises and labelled a “continuing mystery,” the ongoing search for the missing Malaysia Airlines Flight 370 offers no closure. With endless media attention, and negative reactions of stakeholders to every decision made by the airline, this study investigates the types of emotions found in social media posted by publics to the MH370 search suspension announcement. It content analyzed 5,062 real-time tweet messages guided by the revised integrated crisis mapping model. Our findings indicated that, in addition to the four original emotions posited, there was a fifth emotion because of the long-drawn crisis and only two dominant emotions were similar to the model. A redrawn version to better encapsulate all the emotions is offered for one quadrant in the model. Implications for both crisis communication scholarship and the importance of social listening for organizations are discussed.

Do messages spread widely also diffuse fast? Examining the effects of message characteristics on information diffusion

Research examining the information diffusion on social media focused on the size of diffusion but has largely missed the speed by which information spreads over time. To fill this void, this study explores what intrinsic message characteristics drive information not only to reach many people but also to spread fast across social networks. Using a computational social science approach, this study collected all the original tweets from the official Twitter account of the Centers for Disease Control and Prevention (CDC) for 12 months (N = 1934). The diffusion of each tweet message was tracked for six days following its publication. The results showed that tweets with severity, efficacy, and call-for-action information achieved greater diffusion sizes and faster speed. Negative emotional words drove the information to spread widely and quickly, while positive emotional words suppressed the diffusion. Tweet messages with less affiliative words were linked to a larger size and faster speed. The study adds new insights into the dynamic processes of information diffusion and provides practical guidance to promote widespread and fast diffusion on social media.

Twitter message types, health beliefs, and vaccine attitudes during the 2015 measles outbreak in California

Social media not only provide platforms for the public to obtain information about a disease but also allow them to share their opinions and experiences about it. This study analyzed 3000 tweets systematically selected from over 1 million tweets posted during the 2015 California measles outbreak. News updates were the most tweeted messages (41.4%), followed by personal opinions (33.7%), resources (19.4%), personal experiences (2.5%), and questions (1.6%). Susceptibility was the most discussed health belief (21.8%), followed by cues to action (18.9%) and severity (13.0%). Individuals were significantly more likely to discuss severity. Nonprofit organizations were significantly more likely to offer cues to action than other user types, and media were less likely to include cues to action than other user types. Pro-vaccine tweets were more likely to contain links to traditional mainstream media sources such as newspapers and magazines, and anti-vaccine tweets were more likely to link to emerging news websites. Understanding who posts what on social media during an infectious disease outbreak allows public health agencies to better assess the public's attitudes, sentiments, and needs in order to provide timely and effective information.

Analisis Komunikasi Antar Penggemar Seventeen Sebagai Cyberfandom Di Twitter

Research on “The Analysis of Communication Among Fans of Seventeen as a Cyberfandom on Twitter” comes from the observation of the fan phenomenon of a K-Pop group Seventeen on Twitter. Seventeen fans, known as Carat, use Twitter as a medium to communicate and share information between fans. The theory that used in this article are CMC theory, new media theory and cyber community theory. This article used a method of netnography or online ethnography conducted with interviews and online observation of the activitiy of Seventeen fans Twitter account. The analysis technique used in this article is qualitative analysis technique by using data reduction, data presentation and data conclusion. The results of this paper indicates that Seventeen fans exchange information and communicate with each other using Twitter. Carat uses the tweets, reply, quote tweet, retweet and direct message to communicate and exchange information on Twitter.

#CycloneGaja-Rank based Credibility Analysis System in social media during the crisis

Abstract Online Social Networking (OSN) is a platform, which allows users to impart information to other people. Twitter plays an instrumental role in public and political conversations related to the particular crisis events. This research work aims to rank the credibility of the tweets associated with the Natural Disasters and identifies the tweets that spread fake information. The authors propose a Credibility-Analysis-System to rank tweet messages according to their credibility. The authors have compared the performance of tweet credibility with state-of-art of machine learning algorithms such as Naive Bayes classifier, SVM rank algorithm, and Random Forest classifier. The experimentation has been performed on the real-time dataset collected during the Gaja cyclone event. The results present the Random Forest classifier to be the most suitable algorithm for credibility analysis with an accuracy of 87%.

Rumor has it: The effects of virality metrics on rumor believability and transmission on Twitter

This study examines the effects of virality metrics on the judgment of a rumor on Twitter. In an online experiment ( N = 203), a tweet about the rumor related to the issue of food safety was presented with the variations of virality metrics (i.e. the number of retweets, likes, and replies). The results showed that exposure to high virality metrics resulted in greater perceptions of descriptive norms and injunctive norms of sharing the tweet message, and in turn, increased message believability and viral behavioral intentions. Overall, findings demonstrated the normative influence of virality metrics. Implications of the current findings are discussed in light of social norm theories and online rumor literature.

TipMaster: A Knowledge Base of Authoritative Local News Sources on Social Media

Twitter has become an important online source for real-time news dissemination. Especially, official accounts of local government and media outlets have provided newsworthy and authoritative information, revealing local trends and breaking news. In this paper, we describe TipMaster an automatically constructed knowledge base of Twitter accounts that are likely to report local news, from government agencies to local media outlets. First, we implement classifiers for detecting these accounts by integrating heterogeneous information from the accounts' textual metadata, profile images, and their tweet messages. Next, we demonstrate two use cases for TipMaster: 1) as a platform that monitors real-time social media messages for local breaking news, and 2) as an authoritative source for verifying nascent rumors. Experimental results show that our account classification algorithms achieve both high precision and recall (around 90%). The demonstrated case studies prove that our platform is able to detect local breaking news or debunk emergent rumors faster than mainstream media sources.

Few are as Good as Many: An Ontology-Based Tweet Spam Detection Approach

Due to the high popularity of Twitter, spammers tend to favor its use in spreading their commercial messages. In the context of detecting twitter spams, different statistical and behavioral analysis approaches were proposed. However, these techniques suffer from many limitations due to: 1) ongoing changes to Twitter’s streaming API which constrains access to a user’s list of followers/followees; 2) spammer’s creativity in building diverse messages; 3) use of embedded links and new accounts; and 4) need for analyzing different characteristics about users without their consent. To address the aforementioned challenges, we propose a novel ontology-based approach for spam detection over Twitter during events by analyzing the relationship between ham user tweets versus spams. Our approach relies solely on public tweet messages while performing the analysis and classification tasks. In this context, ontologies are derived and used to generate a dictionary that validates real tweet messages from random topics. Similarity ratio among the dictionary and tweets is used to reflect the legitimacy of the messages. Experiments conducted on real tweet data illustrate that message-to-message techniques achieved a low detection rate compared with our ontology-based approach which outperforms them by approximately 200%, in addition to promising scalability for large data analysis.

Extracting Criminal-Related Events from Arabic Tweets

Recently, Twitter as one of social networks has been considered as a rich source of spatio-temporal information and significant revenue for mining data. Event detection from tweets can help to predict more serious real-world events. Such as: criminal events, natural hazards, and the spread of epidemics. Etc. This paper deals with event-based extraction for criminal incidents from Arabic tweets. It presents a framework that supports automated extraction of spatial and temporal information from tweets. The proposed approach is based on combining various indicators, including the names of places and temporal expressions that appear in the tweet message, related tweeting time, and additional locations from the user's profile. The effectiveness of the system was evaluated in term of recall, precision and f-measure.

Posting from beyond the grave: An autopsy of consumer attitudes toward promotional communication in a posthumous context

It is now common for consumers to mourn and commemorate deceased celebrities on social networking sites (SNS). In addition to grieving messages, however, celebrities' posthumous SNS increasingly include promotional content; late celebrities, from Marilyn Monroe to David Bowie, regularly tweet or post promotional messages from the afterlife. The aim of this paper is to explore consumer attitudes toward marketing communication in a posthumous environment. Prior research has consistently documented consumers' rejection of marketing activities in other sacred environments. By investigating the interaction effect of celebrity attachment (fans vs. non-fans) and death salience on celebrities' posthumous SNS, this research is the first to identify a positive response to marketing communication in a sacred environment. Three experimental studies demonstrate that although both fans and non-fans react negatively to promotional messages when death is salient, over time, celebrity attachment diminishes death salience and fosters a positive response of fans toward promotional messages on celebrities' posthumous SNS. Celebrity attachment may favor artificial immortality and, as such, distance the late celebrity from death-related thoughts.

Normative Mechanism of Rumor Dissemination on Twitter.

This study aims to understand the impact of descriptive social norms on message believability and transmission and its underlying mechanism on Twitter. Using two types of information (i.e., news and rumor) presented as a tweet message, the influence of the number of retweets as a normative cue was tested. A result of an online experiment with 639 Twitter users suggests that regardless of the label of the information, message believability and intention to share were stronger for a tweet with a high number of retweets. The mediation test showed that the presumption that a message is believable to others mediates the relationship between a high number of retweets and message believability.

Sharing #Depression in Twitter?: A Content Analysis of Tweets and User Interactions in South Korea and the U.S.

Depression is a common mental disorder that impacts over 350 million people worldwide. South Korea and the U.S. are two countries that statistically, show high rates of depression and suicide. The study compares depression-related tweets from South Korea and the U.S., with the aim of understanding the impact of culture on depression-related tweets. Various content themes, degrees of interactions, levels of sharing patterns, and user characteristics of depressionrelated tweets from both countries were examined. The study findings revealed that for tweet message content, both Korean and US tweets contained individuals’ personal feelings, opinions, interest in depression-related topics, and the encouragement toward those who may suffer from depression and suicide. For user interactivity and sharing activities, US users retweeted more and had more ‘likes’ than Korean users. A correlation analysis revealed that the number of retweets was significantly positively associated with the number of likes in both cultures. Theoretical, methodological, and practical implications are discussed.

잠재 구조적 SVM을 활용한 감성 분석기

본 연구에서는 댓글(음식점/영화/모바일제품) 및 도메인이 없는 트위터 데이터에 대한 감성 분석을 수행하고, 각 문장에 대한 object(or aspect)와 opinion word를 추출하는 시스템을 개발하고 평가한다. 감성 분석을 수행하기 위해 Structural SVM 알고리즘과 Latent Structural SVM 알고리즘을 사용하여 비교 평가하였으며, 실험 결과 Latent Structural SVM이 더 좋은 성능을 보였으며, 구문 분석을 통해 분석된 VP, NP정보를 활용하여 object(aspect)와 opinion word를 추출할 수 있음을 보였다. 또한, 실제 서비스에 활용하기 위해 감성 탐지기를 개발하고 평가하였다.

An automated system for grammatical analysis of Twitter messages. A learning task application

This paper describes an educational study involving the use of Twitter as a way to enhance High School students’ interaction while improving the linguistic quality of their messages. For this purpose, an interactive system has been developed for Twitter collection and analysis from grammatical perspective. The automated system involves a comprehensive data normalization phase, which allows us to identify any unknown token, and a grammatical analysis system. The latter makes use of a logical reasoning on bi-gram token representation as well as a simple rule-based reasoning in case of named-entity detection. The developed system allows the user to perform spatial, topic-based or identity-based search functionalities. Besides, the system generates interrupt to moderator (s) together with some statistical parameters related to user activity as soon as a linguistic inconsistency has been detected in order to take relevant course of actions. The automated system allows us to identify both the text normalization issues and the grammatical inconsistencies. The latter makes use of logical reasoning using bi-gram Wikipedia matching. A statistical analysis of tweet messages gathered from students that took part to this study has been carried out. Besides, the contribution of the peers to the improvement of the linguistic quality of users’ messages has been quantified and investigated. The study demonstrates the interest of the participants to this new learning experience and evaluates the influence of the peers on their writing skills. Especially, the visibility and noticeability of Twitter messages to a large audience have been found to contribute widely to raise students’ awareness about the linguistic quality of their messages. The study has also revealed the predominance of the slang language in their daily Twitter writings. Such abbreviations have shown to pose the greatest challenge for any automatic text analysis. Similarly named-entity identification and handling have also been shown to be very challenging, especially, given the nature of Twitter messages where capitalizing is often employed for emphasize as well.

Revisiting the Task Media Fit Model in the era of Web 2.0: Twitter use and interaction in the healthcare sector

Advances in technology have led to the development of social media and subsequently new channels of communication. This paper refines the established Task Media Fit Model in light of such changes, using business marketers' use of Twitter and followers' responses to tweeted messages for this preliminary conceptual development. Results show that business marketers use different embedded media according to the function of a tweet message. Follower responses to those messages do not vary with the task performed by the tweet, while responses differ with the type of embedded link. Findings from this investigation are used to develop a modified version of the Task Media Fit Model specifically for Twitter.

User and Content Characteristics of Public Tweets Referencing Little Cigars.

Compared to cigarettes, little cigars and cigarillos (LCC) are minimally regulated, affordable, and widely available to young people. Because Twitter is a preferred mode of communication among younger people, product portrayals may be useful for informing both interventions and public health or tobacco policy. A mixed-methods study was implemented to analyze the content of public tweets (N = 288) and profile photos sampled from a search of 2 LCC brands (Black & Mild and Swisher Sweets). Metadata and manifest attributes of profile photo demographic features and tweet message features were coded and analyzed. Thematic analysis of the tweets was conducted using an iterative immersion/ crystallization method. Tweeters were most often boys or men (63%) and appeared young (76%). Prevalent content themes included expressing affiliation for the LCC product and reporting smoking activity. Although men and women tweeted affiliation for LCC products and reported smoking activity in similar numbers, women were significantly less likely to tweet about blunting than men. Twitter provides a potentially potent source of nuanced information about how young people are using little cigars. These observed characteristics may be useful to inform counter-messaging strategies and interventions.

Conversations and Campaign Dynamics in a Hybrid Media Environment: Use of Twitter by Members of the German Bundestag

This article examines how Members of the German Bundestag (MdBs) used Twitter in the context of the country’s 2013 federal elections. In particular, we explore the dynamics in the MdBs’ use of Twitter during different periods of the electoral term: How do the tweeting habits of MdBs differ by party before and during the election campaign in (a) public versus personal communication and (b) campaign versus policy messages? How are the selection of interaction partners, centralization on leading actors, and reciprocity of the MdBs’ Twitter networks affected by election campaigning? We address these questions by conducting a content analysis combined with a network analysis of interaction patterns. The comparative application of both methods explains the differences of MdBs’ networks. The comparison clearly exhibits election campaign-driven changes related to the amount of activity and the character of tweeted messages. During the campaign period, MdBs’ tweets clearly discussed specific policies less than before. Tweeting about one’s personal life occurred also less frequently in the final campaign stage. Instead, the MdBs mainly complement other forms of election campaigning through a vivid metacommunication on campaign developments. Network relations reflect these variations and were less often reciprocated in proximity to the election and showed a higher degree of group homophily. We also found a substantial representation of print and broadcast media actors in the examined @reply networks. It is likely that these interactions and conversations with journalists are part of an MdB’s individual performance of “news management.”

Context-aware Models for Twitter Sentiment Analysis

Recent works on Sentiment Analysis over Twitter are tied to the idea that the sentiment can be completely captured after reading an incoming tweet. However, tweets are filtered through streams of posts, so that a wider context, e.g. a topic, is always available. In this work, the contribution of this contextual information is investigated for the detection of the polarity of tweet messages. We modeled the polarity detection problem as a sequential classification task over streams of tweets. A Markovian formulation of the Support Vector Machine discriminative model has been here adopted to assign the sentiment polarity to entire sequences. The experimental evaluation proves that sequential tagging better embodies evidence about the contexts and is able to increase the accuracy of the resulting polarity detection process. These evidences are strengthened as experiments are successfully carried out over two different languages: Italian and English. Results are particularly interesting as the approach is flexible and does not rely on any manually coded resources.