Real-time Tweets Research Articles

Predicting Geolocation of Tweets: Using Combination of CNN and BiLSTM

Twitter is one of the most popular micro-blogging and social networking platforms where users post their opinions, preferences, activities, thoughts, views, etc., in form of tweets within the limit of 280 characters. In order to study and analyse the social behavior and activities of a user across a region, it becomes necessary to identify the location of the tweet. This paper aims to predict geolocation of real-time tweets at the city level collected for a period of 30 days by using a combination of convolutional neural network and a bidirectional long short-term memory by extracting features within the tweets and features associated with the tweets. We have also compared our results with previous baseline models and the findings of our experiment show a significant improvement over baselines methods achieving an accuracy of 92.6 with a median error of 22.4 km at city level prediction.

COVID-19 Vaccine Hesitancy in Canada: Content Analysis of Tweets Using the Theoretical Domains Framework.

BackgroundWith the approval of two COVID-19 vaccines in Canada, many people feel a sense of relief, as hope is on the horizon. However, only about 75% of people in Canada plan to receive one of the vaccines.ObjectiveThe purpose of this study is to determine the reasons why people in Canada feel hesitant toward receiving a COVID-19 vaccine.MethodsWe screened 3915 tweets from public Twitter profiles in Canada by using the search words “vaccine” and “COVID.” The tweets that met the inclusion criteria (ie, those about COVID-19 vaccine hesitancy) were coded via content analysis. Codes were then organized into themes and interpreted by using the Theoretical Domains Framework.ResultsOverall, 605 tweets were identified as those about COVID-19 vaccine hesitancy. Vaccine hesitancy stemmed from the following themes: concerns over safety, suspicion about political or economic forces driving the COVID-19 pandemic or vaccine development, a lack of knowledge about the vaccine, antivaccine or confusing messages from authority figures, and a lack of legal liability from vaccine companies. This study also examined mistrust toward the medical industry not due to hesitancy, but due to the legacy of communities marginalized by health care institutions. These themes were categorized into the following five Theoretical Domains Framework constructs: knowledge, beliefs about consequences, environmental context and resources, social influence, and emotion.ConclusionsWith the World Health Organization stating that one of the worst threats to global health is vaccine hesitancy, it is important to have a comprehensive understanding of the reasons behind this reluctance. By using a behavioral science framework, this study adds to the emerging knowledge about vaccine hesitancy in relation to COVID-19 vaccines by analyzing public discourse in tweets in real time. Health care leaders and clinicians may use this knowledge to develop public health interventions that are responsive to the concerns of people who are hesitant to receive vaccines.

Analysing Twitter Data for Phishing Tweets Identification

Detecting threats like adult, violent, and phishing tweets on online social networks is a crucial issue in recent years. The aim of the work is to identify phishing content from the users' perspective in real-time tweets. To outline such content comprehensively, lexicon analysis with sentiments are encapsulated to investigate tweets that yield phishing dynamic keywords, while some features and parameters are altered to optimize the performance. To support the preliminary study, the approach is rigorously designed to assemble users' opinions on completely different classes of phishing content. Each direct and indirect opinions as well as recently projected opinions are listed to characterize all sorts of phishing content. The authors use word level analysis with sentiments to build keyword blacklist lexicons. High promising results and high level of accuracy and performance are obtained experimentally if compared with the alternative algorithms.

A probabilistic clustering model for hate speech classification in twitter

The key challenges for automatic hate-speech classification in Twitter are the lack of generic architecture, imprecision, threshold settings and fragmentation issues. Most studies used binary classifiers for hate speech classification, but these classifiers cannot really capture other emotions that may overlap between positive or negative class. Hence, a probabilistic clustering model for hate speech classification in twitter was developed to tackle problems with hate speech classification. A metadata extractor was used to collect tweets containing hate speech keywords and a crowd-sourced experts was employed to label the collected hate tweets into two categories: hate speech and non-hate speech. Features representation was done with Term Frequency- Inverse Document Frequency (TF-IDF) model and enhanced with topics inferred by a Bayes classifier. A rule-based clustering method was used to automatically classify real-time tweets into the correct topic clusters. Fuzzy logic was then used for hate speech classification using semantic fuzzy rules and a score computation module. From the evaluation results, it was observed that the developed model performed better in hate speech detection with F1-sore of 0.9256 using a 5-fold cross validation. Similarly, the developed model for hate speech classification performed better with F1-score of 91.5 compared to related models. The developed model also indicates a more perfect test having an AUC of 0.9645, when compared to similar methods. The Paired Sample t-Test validated the efficiency of the developed model for hate speech classification.

An Emotion Care Model using Multimodal Textual Analysis on COVID-19.

At the dawn of the year 2020, the world was hit by a significant pandemic COVID-19, that traumatized the entire planet. The infectious spread grew in leaps and bounds and forced the policymakers and governments to move towards lockdown. The lockdown further compelled people to stay under house arrest, which further resulted in an outbreak of emotions on social media platforms. Perceiving people's emotional state during these times becomes critically and strategically important for the government and the policymakers. In this regard, a novel emotion care scheme has been proposed in this paper to analyze multimodal textual data contained in real-time tweets related to COVID-19. Moreover, this paper studies 8-scale emotions (Anger, Anticipation, Disgust, Fear, Joy, Sadness, Surprise, and Trust) over multiple categories such as nature, lockdown, health, education, market, and politics. This is the first of its kind linguistic analysis on multiple modes pertaining to the pandemic to the best of our understanding. Taking India as a case study, we inferred from this textual analysis that ‘joy’ has been lesser towards everything (~9-15%) but nature (~17%) due to the apparent fact of lessened pollution. The education system entailed more trust (~29%) due to teachers' fraternity's consistent efforts. The health sector witnessed sadness (~16%) and fear (~18%) as the dominant emotions among the masses as human lives were at stake. Additionally, the state-wise and emotion-wise depiction is also provided. An interactive internet application has also been developed for the same.

Multi-contextual spammer detection for online social networks

With prevalent usage of internetworking, online social platforms have greatly revolutionized the traditional exchange of information. Social networking sites, including Facebook, Twitter, LinkedIn, etc., are being extensively used by the masses to create social as well as professional content and connections. However, due to wide transparency in information spread, the user profiles often remain vulnerable to spammers. Such threats occur in terms of harmful links being posted, involving eavesdropping attack and making fake identities of other people. The proposed model performs detection of spammers on the basis of real-time tweets extracted from Twitter. Several features have been taken into consideration while detecting the spammed Twitter profiles. The social network features include replies, mentions, web links, trending tags and profile reputation. Another feature has been introduced in the content-based feature to analyze sentiment of each tweet with consideration of the emoticons. Further, machine learning algorithms have been applied to conduct performance evaluation with respect to accuracy and spam accounts detection.

Using community information for natural disaster alerts

Recently, the ceaseless rise in the global average temperature has led to extreme climates in which natural disasters, such as droughts, hurricanes, earthquakes and floods, are becoming increasingly serious. Recent research has found that social media typically reflects disasters earlier than official communication channels. In this study, the idea of collecting information on flood disasters caused during the periods of typhoons and heavy rains for a city from the plain text messages released by social media by means of a term frequency (TF) and sliding window approach is proposed. The dataset analysed here contains a total of 292 articles and 12,484 tweets. This research determines how to establish a warning mechanism, with an added notification time for flooding disasters, and it shows how to provide relevant disaster relief personnel with references. This article contributes by combining social media data with emergency management information cloud (EMIC) data, especially in the context of having a mechanism for warning about flooding disasters. According to the experimental results, a sliding window of 90 min and a sliding gap of 10 min obtained the best F-measure value ( F = 0.315). The event studied was Typhoon Megi (September 2016), which caused major flooding in Tainan. For the Typhoon Megi event, the flood disaster location database had 161 streets available for matching. Based on the experimental results, it is possible to obtain a high-precision (90% or higher) accuracy rate from real-time tweet data by exploiting a social media dataset.

A Framework for Analyzing Real-Time Tweets to Detect Terrorist Activities

A Framework for Analyzing Real-Time Tweets to Detect Terrorist Activities

Social networking mood recognition algorithm for conflict detection and management of Indian educational institutions

Conflict detection and management is a much-needed strategy in educational institutions nowadays. The frequency of clashes, protests, strikes and agitations is rising at an alarming level, particularly with the extensive usage of information and communication technology. This unprecedented issue has cobblestone the research problem and the objective of this study. In order to ensure detection of conflicts using a modified Naive Bayes algorithm that would assess the sentiments and mood recognition from the tweets of the stakeholders with respect to trends leading to clashes, protests, strikes and agitations existing within the educational environment. Consequently, continuous monitoring and surveillance of the social networking platform have been made to understand the mood recognition and assessment from real-time tweets published on Twitter that would be streamed through Twitter Application Program Interface over the user-given text that would serve as an input for developing early conflict detection strategies. Moreover, Stanford parser has been used to extract keywords that would subsequently be deployed over the tweets for the mood detection module to evaluate the keywords using the preprocessed dictionary. Here, the mood detection module extracts various mood states from the given chorus that supplements the current prevailing sentiments of the stakeholders in academic institutions. This extraction would result in devising strategic formulation and implementation of remedial measures for resolving the unrest in an environment. This would culminate in a healthy educational environment that would ultimately result in overall growth, development, harmony and prosperity of the institutions.

Influence of Social Media Analytics on Online Food Delivery Systems

Online food delivery applications have gained significant attention in the metropolitan cities by diminishing the burden of traveling and waiting time by offering online food delivery options for various dishes from many such restaurants. Users enjoy these services and share their experiences and opinions on social media platforms that impact the trust of customers and change their purchasing habits. This drastic revolution of user activities is an opportunity for targeted social marketing. This research is based on Twitter's data and aimed to identify the influence of social media in food delivery e-commerce businesses including decision making, marketing strategy, consumer behavior analysis, and improving brand reputation. In this article, the authors proposed an Apache Spark-based social media analytics framework to process the tweets in real time to identify the influences of generated insights on e-commerce decision making. The experimental analysis highlighted the exponentially grown influence of social media in food delivery e-commerce portals in past years.

Near real-time twitter spam detection with machine learning techniques

ABSTRACT The popularity of social media networks, such as Twitter, leads to an increasing number of spamming activities. Researchers employed various machine learning methods to detect Twitter spam. However, majorities of existing researches are limited to theoretically study, few of them can apply detection techniques to real-time scenario. In this paper, we bridge the gap by proposing a near real-time Twitter spam detection system, which provides near real-time tweets data acquisition, light-weight features extraction from a specific Twitter account, training detection model, and online visualizing detection results. In this system, account-based and content-based features are extracted to facilitate spam detection. The models that are applied to our Twitter spam detection system are trained based on 1.5 million public tweets and nine mainstream algorithms. In addition, in order to efficiently reduce training time spent on massive data and save the cost of model updating, a parallel computing technique is introduced to train and update the models in this system. Empirical results verify that the model can achieve satisfactory performance based on our datasets. Furthermore, we implement a near real-time Twitter spam detection system which can better protect users from combating spams. This system also acts as a tweets collection tool, allowing researchers to test the performance of trained classifiers in realistic scenarios.

Twitter metrics complement traditional conference evaluations to evaluate knowledge translation at a National Emergency Medicine Conference.

Conferences are designed for knowledge translation, but traditional conference evaluations are inadequate. We lack studies that explore alternative metrics to traditional evaluation metrics. We sought to determine how traditional evaluation metrics and Twitter metrics performed using data from a conference of the Canadian Association of Emergency Physicians (CAEP). This study used a retrospective design to compare social media posts and tradition evaluations related to an annual specialty conference. A post ("tweet") on the social media platform Twitter was included if it associated with a session. We differentiated original and discussion tweets from retweets. We weighted the numbers of tweets and retweets to comprise a novel Twitter Discussion Index. We extracted the speaker score from the conference evaluation. We performed descriptive statistics and correlation analyses. Of a total of 3,804 tweets, 2,218 (58.3%) were session-specific. Forty-eight percent (48%) of all sessions received tweets (mean = 11.7 tweets; 95% CI of 0 to 57.5; range, 0-401), with a median Twitter Discussion Index score of 8 (interquartile range, 0 to 27). In the 111 standard presentations, 85 had traditional evaluation metrics and 71 received tweets (p > 0.05), while 57 received both. Twenty (20 of 71; 28%) moderated posters and 44% (40 of 92) posters or oral abstracts received tweets without traditional evaluation metrics. We found no significant correlation between Twitter Discussion Index and traditional evaluation metrics (R = 0.087). We found no correlation between traditional evaluation metrics and Twitter metrics. However, in many sessions with and without traditional evaluation metrics, audience created real-time tweets to disseminate knowledge. Future conference organizers could use Twitter metrics as a complement to traditional evaluation metrics to evaluate knowledge translation and dissemination.

Beyond the hashtag – An exploration of tweeting and replies at the European Society of Surgical Oncology 39th clinical conference (ESSO39)

IntroductionPrevious studies of conference tweeting have focused on tweets that used the conference hashtag. The aim of this study was to document responses beyond a specific conference hashtag. MethodsObservational study exploring replies to tweets for the 39th meeting of ESSO (ESSO39), Rotterdam, 9–11 October 2019. An extract of #ESSO39 tweets was obtained using NodeXL. Replies to these tweets were identified by viewing the tweets via Twitter.com. ResultsThere were 210 tweets posted using the #ESSO39 hashtag by 64 tweeters. However, extending the analysis to include responses that did not use the hashtag, tweets using the hashtag only represent 54% of all tweets posted or quoted at the conference, and only 49% of the tweeters posting content or quoted in tweets. Based on this study of ESSO39 therefore roughly half of tweets and contributors to conference tweeting were not captured by focusing simply on the conference hashtag (#ESSO39). Mentioning another tweeter(s) in a tweet or response was associated with more retweets, as was including the hashtag in replies. ConclusionsTwitter activity at medical conferences extends beyond the conference hashtag. Almost half of the tweeting was beyond the hashtag. To increase visibility of tweets, conference delegates should include the conference hashtag and mention other tweeters in their tweets and responses. Searching for tweets is an active process requiring users to click into replies. Twitter and third-party social media tools should improve identification and display of responses, showing the branching structure of replies and quoting tweets in real time.

Prototyping a Social Media Flooding Photo Screening System Based on Deep Learning

This article aims to implement a prototype screening system to identify flooding-related photos from social media. These photos, associated with their geographic locations, can provide free, timely, and reliable visual information about flood events to the decision-makers. This screening system, designed for application to social media images, includes several key modules: tweet/image downloading, flooding photo detection, and a WebGIS application for human verification. In this study, a training dataset of 4800 flooding photos was built based on an iterative method using a convolutional neural network (CNN) developed and trained to detect flooding photos. The system was designed in a way that the CNN can be re-trained by a larger training dataset when more analyst-verified flooding photos are being added to the training set in an iterative manner. The total accuracy of flooding photo detection was 93% in a balanced test set, and the precision ranges from 46–63% in the highly imbalanced real-time tweets. The system is plug-in enabled, permitting flexible changes to the classification module. Therefore, the system architecture and key components may be utilized in other types of disaster events, such as wildfires, earthquakes for the damage/impact assessment.

Regional Influenza Prediction with Sampling Twitter Data and PDE Model.

The large volume of geotagged Twitter streaming data on flu epidemics provides chances for researchers to explore, model, and predict the trends of flu cases in a timely manner. However, the explosive growth of data from social media makes data sampling a natural choice. In this paper, we develop a method for influenza prediction based on the real-time tweet data from social media, and this method ensures real-time prediction and is applicable to sampling data. Specifically, we first simulate the sampling process of flu tweets, and then develop a specific partial differential equation (PDE) model to characterize and predict the aggregated flu tweet volumes. Our PDE model incorporates the effects of flu spreading, flu recovery, and active human interventions for reducing flu. Our extensive simulation results show that this PDE model can almost eliminate the data reduction effects from the sampling process: It requires lesser historical data but achieves stronger prediction results with a relative accuracy of over 90% on the 1% sampling data. Even for the more aggressive data sampling ratios such as 0.1% and 0.01% sampling, our model is still able to achieve relative accuracies of 85% and 83%, respectively. These promising results highlight the ability of our mechanistic PDE model in predicting temporal–spatial patterns of flu trends even in the scenario of small sampling Twitter data.

Extraction Opinion of Social Media in Higher Education Using Sentiment Analysis

The purpose of this research is to extract social media Twitter opinion on a tertiary institution using sentiment analysis. The results of sentiment analysis will provide input to universities as a form of evaluation of management performance in managing institutions. Sentiment analysis generated using the Naïve Bayes Classifier method which is classified into 4 classes: positive, normal, negative and unknown. This study uses 1000 data tweets used for training data needs. The data is classified manually to determine the sentiment of the tweet. Then 20 tweet data is used for testing.
 The results of this study produce a system that can classify sentiments automatically with 75% test results for sentiment, some obstacles in processing real-time tweets such as duplicate tweets (spam tweets), Indonesian structures that are quite complex and diverse.

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.

A semantic graph-based keyword extraction model using ranking method on big social data

Identification of an influential node is essential to control Online Social Networks. It is a crucial task with various kinds of real-time usages such as information retrieval and recommendation, instinctive keyword indexing, viral marketing, instinctive classification, and filtering. Existing graph-based methods consider only numeric measures such as centrality, degree, and betweenness for keyword extraction. In this research, a relatively effective method called Semantic graph-based Keyword Extraction Method (SKEM) from Twitter using ranking methods is proposed. In the proposed model, the exhaustive preprocessing is carried out, and a semantic graph-based model has been constructed. The numeric graph metrics are then used to weigh the nodes of the semantic graph. Page Rank algorithm is applied to arrange the nodes, and the top ten nodes that are found to be very relevant and effectively represent the most influential node. Combining both semantic as well as numeric graph metrics, it greatly enhances the quality of keywords extracted. The extensive preprocessing enhances the quality of the input and minimizes the noise in the input. The keywords extracted by the proposed model have been more relevant and meaningful. The performance of the proposed SKEM model is validated with real-time tweets of Twitter API. The experimental results are confirmed that the proposed method is achieving high performance in terms of precision, recall, and F-measure.

Data Mining and Machine Learning: Design a Generalized Real Time Sentiment Analysis System on Tweeter Data Using Natural Language Processing

Sentiment analysis is a task, that is becoming recently important for numerous companies. Because the consigner subscriptions on social media like Facebook, twitter and other side get their product reviews. If the company wants to track tweets about their brand to command over the impact on time or many website analyze the comments on their articles. This will help them to track comments and impact. So the sentiment analysis is an automated system that collects and analyzes the content and generates the desired results. This paper proposes a sentiment analysis system for twitter posts. Proposed system will work on real time tweets. System is also designed in such a way that this can analyze data related to any topic. Python programming language is used to extract tweets form twitter feeds. Proposed system also calculates the level of sentiments. That how much negative or positive tweets are. This paper also presents some real time result analysis.

Statistical Analysis for Twitter Spam Detection

The spreading and learning of new discoveries and information is made available using current online social networks. In Recent days, the solutions may be irrelevant to the actual content; also termed as attacks in the layman’s term such attacks are been performed on Twitter as well and called as Twitter spammers. The quality of data is being compromised by addition of malicious and harmful information using URL, bio, emoticons, audio, images/videos & hash-tags through different accounts by exchanging tweets, personal messages (Direct Message’s) & re-tweets. Misleading sites may be linked with the malicious links which may affect adverse effects on the user and also interfere in their decision making processes. To improve user-experience from the spammers attacks, the training twitter dataset are applied and then by extracting and using the 12 lightweight features like user’s age, number of followers, count of tweets and re-tweets, etc. are used to distinguish the spam from non-spam. For enhancing the performance, the discretization of the function is important for transmission of spam detection between tweets. Our system creates classification model for Spam detection which includes binary classification and automatic learning algorithms viz. Naïve Bayes classifier or Support Vector Machine classifier which understands the behaviour of the model. The system will categorize the tweets from datasets into Spam and Non-spam classes and provide the user’s feed with only the relevant information. The system will report the impact of data-related factors such as relationship between spam and non-spam tweets, size of training dataset, data sampling and detection performance. The proposed system’s function is detection and analysis of the simple and variable twitter spam over time. The spam detection is a major challenge for the system and shortens the gap between performance appraisals and focuses primarily on data, features and patterns to identify real user and informing it about the spam tweets along with the performance statistics. The work is to detect spammed tweets in real time, since the new tweets may show patterns and this will help for training and updating dataset and in knowledge base.