Online Social Networking Platforms Research Articles

User Profile Construction Based on High-Dimensional Features Extracted by Stacking Ensemble Learning

Online social networks, as platforms for personal expression, have evolved into complex networks integrating political and social dimensions. This evolution has shifted the focus of network governance from addressing hacking activities to mitigating unpredictable social behaviors, such as the malicious manipulation of public opinion, the doxing of ordinary users, and cyberbullying. However, the sparsity of data and the concealed nature of user behavior pose significant challenges to existing network reconnaissance technologies. In this study, we focus on constructing user profiles on online social network platforms by extracting features to build deep user profiles based on behavioral patterns. Drawing inspiration from the 5Cs principle of credit evaluation, we refine it into a 3Cs principle tailored for user profiling on social network platforms and associate it with user behavioral patterns. To further analyze user behavior, a high-dimensional feature extraction method is proposed using an improved stacking ensemble learning model. Based on experimental data analysis, the most suitable base algorithms for high-dimensional feature extraction are identified. Experimental results demonstrate that the integration of high-dimensional features improved the behavior prediction accuracy of the profiling model by 9.26% on balanced datasets and enhanced the AUC (area under the curve) metric by 3.69% on imbalanced datasets. The proposed method effectively increases the depth and generalization performance of user profiling.

Effects of network connectivity and functional diversity distribution on human collective ideation

Human collective tasks in teams and organizations increasingly require participation of members with diverse backgrounds working in networked social environments. However, little is known about how network structure and the functional diversity of member backgrounds would interact with each other and affect collective processes. Here we conducted three sets of human-subject experiments which involved 617 university students who collaborated anonymously in a collective ideation task on a custom-made online social network platform. We found that spatially clustered collectives with assortative background distribution tended to explore more diverse ideas than in other conditions, whereas collectives with random background distribution consistently generated ideas with the highest utility. We also found that higher network connectivity may improve individuals’ overall experience but may not improve the collective performance regarding idea generation, idea diversity, and final idea quality.

Relieving Overexposure in Information Diffusion Through a Budget Multi-stage Allocation

When information dissemination campaigns on Online Social Networking platforms are too aggressive, this can easily cause information overexposure. Overexposure can break through the psychological and physiological limits that the audience can tolerate, making it difficult for the audience to obtain reasonable cognitive concepts. For example, the overexposure of information referring to an object in a promotional campaign can lead to inflated expectations in individuals. Building on this, we introduce two indicators for individuals’ expectations and actual utility of the object and design a multi-stage triggered mechanism for seed individuals to explore the relieving overexposure problem in information diffusion. We build a multi-stage information diffusion model and characterize the evolution of individual expectations. We verify the hardness result of the relieving overexposure problem by budget multi-stage allocation, and the non-submodularity and non-monotonicity of the objective function. Addressing the non-monotonic and non-submodular set function, we provide a direct influence-oriented algorithm with a greedy approach. Extensive experiments are performed on four real networks to explore how model parameters and network properties affect the effects of multi-stage triggered strategies for seed individuals. Using the experiments, we found that the seed individual multi-stage incremental triggered strategy of dissemination campaign of information referring to an object shows better performance, and the lower the actual utility of the specific object, the more accurate the promotion strategy needs to be developed.

What target language use characterizes high-performing L2 learners in a migration context?

Aims and Objectives/Purpose/Research Questions: Previous studies have investigated if quantity of target language (TL) use predicts second language (L2) learning. The aim of this study is to understand what kind of TL use promotes host language proficiency among adults in the context of migration, operationalized as productive collocation knowledge. Design/Methodology/Approach: The study takes a cross-sectional quantitative approach to explore what clusters of TL activities predict language proficiency among 63 first language (L1) French L2 Swedish long-term residents in Sweden. Data and Analysis: Data were collected within a larger project. Productive collocation knowledge was assessed by a collocation test and participants’ TL use through the Language Engagement Questionnaire. We conducted a factor analysis to identify patterns of TL use. We then used structural equation modeling to investigate the relationship between TL use and productive collocation knowledge. Findings/Conclusions: Two factors emerged from the factor analysis: “Everyday Interactive Target Language Use” and “Non-interactive Target Language Use.” Only the former was related to productive collocation knowledge as evidenced by the structural equation modeling. “Everyday Interactive Target Language Use” included using the TL at the workplace, in conversations and service encounters, through text messages and emails, participating in online social network platforms, and browsing the internet. Originality: The study explores what specific kind of TL use promotes second language proficiency among adults in a migration context. This is a topic that has not received much attention in previous research. Significance/Implications: By showing that proficient L2 users engage daily with the TL through, for example, social interactions and language use at work, the study points toward the importance of a certain level of socio-economic as well as socio-cultural integration for efficient second language learning among adults in a migration context.

Privacy-Preserving and Diversity-Aware Trust-based Team Formation in Online Social Networks

As online social networks (OSNs) become more prevalent, a new paradigm for problem-solving through crowd-sourcing has emerged. By leveraging the OSN platforms, users can post a problem to be solved and then form a team to collaborate and solve the problem. A common concern in OSNs is how to form effective collaborative teams, as various tasks are completed through online collaborative networks. A team’s diversity in expertise has received high attention to producing high team performance in developing team formation (TF) algorithms. However, the effect of team diversity on performance under different types of tasks has not been extensively studied. Another important issue is how to balance the need to preserve individuals’ privacy with the need to maximize performance through active collaboration, as these two goals may conflict with each other. This research has not been actively studied in the literature. In this work, we develop a TF algorithm in the context of OSNs that can maximize team performance and preserve team members’ privacy under different types of tasks. Our proposed PRivAcy-Diversity-Aware TF framework, called PRADA-TF , is based on trust relationships between users in OSNs where trust is measured based on a user’s expertise and privacy preference levels. The PRADA-TF algorithm considers the team members’ domain expertise, privacy preferences, and the team’s expertise diversity in the process of TF. Our approach employs game-theoretic principles Mechanism Design to motivate self-interested individuals within a TF context, positioning the mechanism designer as the pivotal team leader responsible for assembling the team. We use two real-world datasets (i.e., Netscience and IMDb) to generate different semi-synthetic datasets for constructing trust networks using a belief model (i.e., Subjective Logic) and identifying trustworthy users as candidate team members. We evaluate the effectiveness of our proposed PRADA-TF scheme in four variants against three baseline methods in the literature. Our analysis focuses on three performance metrics for studying OSNs: social welfare, privacy loss, and team diversity.

Heterogeneous Social Network Location Aware Multi Party Data Protection and Privacy System

Social networks are large online communities that allow users to interact and share information. They can be used to connect with friends, family, or potential romantic partners, build a business, or even find a job. The evolution of social media has led to a trend of posting daily photos on online Social Network Platforms (SNPs). The privacy of online photos is often protected carefully by security mechanisms. However, these mechanisms will lose effectiveness when someone spreads the photos to other platforms. This project proposes a blockchain based secure photo sharing framework that provides powerful dissemination control for heterogeneous social network photo sharing. In contrast to security mechanisms running separately in centralized servers that do not trust each other, this framework achieves consistent consensus on photo dissemination control through carefully designed smart contract-based protocols

Threats on online social network platforms: classification, detection, and prevention techniques

Threats on online social network platforms: classification, detection, and prevention techniques

A Federated Social Recommendation Approach with Enhanced Hypergraph Neural Network

In recent years, the development of online social network platforms has led to increased research efforts in social recommendation systems. Unlike traditional recommendation systems, social recommendation systems utilize both user-item interactions and user-user social relations to recommend relevant items, taking into account social homophily and social influence. Graph neural network (GNN) based social recommendation methods have been proposed to model these item interactions and social relations effectively. However, existing GNN-based methods rely on centralized training, which raises privacy concerns and faces challenges in data collection due to regulations and privacy restrictions. Federated learning has emerged as a privacy-preserving alternative. Combining federated learning with GNN-based methods for social recommendation can leverage their respective advantages, but it also introduces new challenges: 1) existing federated recommendation systems often lack the capability to process heterogeneous data, such as user-item interactions and social relations; 2) due to the sparsity of data distributed across different clients, capturing the higher-order relationship information among users becomes challenging and is often overlooked by most federated recommendation systems. To overcome these challenges, we propose a federated social recommendation approach with enhanced hypergraph neural network. We introduce hypergraph graph neural networks (HGNN) to learn user and item embeddings in federated recommendation systems, leveraging the hypergraph structure to address the heterogeneity of data. Based on carefully crafted triangular motifs, we merge user and item nodes to construct hypergraphs on local clients, capturing specific triangular relations. Multiple HGNN channels are used to encode different categories of high-order relations, and an attention mechanism is applied to aggregate the embedded information from these channels. Our experiments on real-world social recommendation datasets demonstrate the effectiveness of the proposed approach. Extensive experiment results on three publicly available datasets validate the effectiveness of the proposed method.

Modeling the dissemination of privacy information in online social networks

AbstractDriven by the rapid development of information technology, online social networks (OSNs) have experienced a fast development in recent years, allowing increasingly more people to share and spread information over OSNs. The rapid rise of OSN platforms such as Facebook and Twitter is sufficient evidence of such development. As one type of information, privacy information can also be created and disseminated over an OSN, posing a severe threat to individual privacy. This article attempts to construct a model for disseminating privacy information in OSNs and to analyze the model by simulating the dissemination process of privacy information in OSNs. First, we establish network models that exhibit the main characteristics of OSNs. Second, by considering the factors related to social relationships, especially intimacy between users and the attention of users to the privacy subject, we derive the parameters for privacy information dissemination models in OSNs. Third, based on the theory of information dissemination dynamics, we construct a model for information dissemination that conforms to the properties of privacy information. We also present some experimental results based on the constructed model and analyze the characteristics of privacy information dissemination. Fourth, we study and verify the various properties of the model through a set of experiments. The proposed model provides the opportunity to better understand the dynamics of privacy information dissemination in OSNs and the effect of user behavior on dissemination.

Detection of Fake Profiles on Online Social Network Platforms: Performance Evaluation of Artificial Intelligence Techniques

Detection of Fake Profiles on Online Social Network Platforms: Performance Evaluation of Artificial Intelligence Techniques

A novel Cosine-Cosine chaotic map-based video encryption scheme

The surge in online activities has led to the increasing popularity of sharing video data across diverse applications, including online education tutorials, social networking, video calling, and OTT platforms. Encryption prevents unauthorized access to the transmitted data over unreliable channels. The well-known features of chaos theory such as random behaviour, unpredictability, and initial parameters dependency facilitate its use in cryptography. Many security issues are faced by chaos-based cryptosystems because of their less complexity. Hence, a new Cosine-Cosine chaotic map characterized by intricate chaotic behaviour is designed in the current study. Additionally, we formulate an original video encryption scheme employing this Cosine-Cosine chaotic map. The encryption process involves five steps, beginning with the segmentation of the original video into frames based on its frame rate. In the second phase, a 384 bits pseudorandom key is generated that is further divided into three subkeys of 128 bits each. The novel Cosine-Cosine chaotic map-based sequence is generated. In the fourth step, red, green, and blue components are encrypted using the pseudorandom key and the chaotic sequence. In the last step, we combine encrypted frames to get cipher video. The security analysis validates that the proposed encryption protects against eavesdropping.

Game-theoretic modeling and analysis of cyberbullying spreading on OSNs

The rapid development of online social networks (OSNs) has enabled seamless cyberbullying activities, which not only cause psychological harm to victims but also bring disharmony to society. From the perspective of OSN platform administrators, strategic decision making towards when/how to launch an anti-cyberbullying campaign is the key to mitigating its negative impact. Hence, this paper develops a dynamic anti-cyberbullying investment solution to mitigate the adverse impact of strategic cyberbullying phenomena cost-effectively. Specifically, this paper first quantifies the payoffs of the attacker/bullying instigator and the defender/platform administrator based on a newly proposed propagation model. Then, considering the budget-constrained factor for both sides, a differential game-theoretical model is proposed to simulate each party's optimal trade-off between cost and benefit. Next, by utilizing Pontryagin Maximum/Minimum Principle, a set of necessary conditions for a Nash equilibrium strategy profile have been proposed. Based on the forward-backward sweep method, an iterative algorithm (SCR optimization algorithm) is tailored to solve the conditions numerically. Through experiments conducted under Facebook, Youtube and Twitter sub-networks, the algorithm's performance has been evaluated by inspecting its convergence speed and payoff compared with randomized profiles. Consequently, the resulting strategies are verified to exhibit superior performance in dealing with cyberbullying spreading.

Toxic Fake News Detection and Classification for Combating COVID-19 Misinformation

The emergence of COVID-19 has led to a surge in fake news on social media, with toxic fake news having adverse effects on individuals, society, and governments. Detecting toxic fake news is crucial, but little prior research has been done in this area. This study aims to address this gap and identify toxic fake news to save time spent on examining nontoxic fake news. To achieve this, multiple datasets were collected from different online social networking platforms such as Facebook and Twitter. The latest samples were obtained by collecting data based on the topmost keywords extracted from the existing datasets. The instances were then labeled as toxic/nontoxic using toxicity analysis, and traditional machine-learning (ML) techniques such as linear support vector machine (SVM), conventional random forest (RF), and transformer-based ML techniques such as bidirectional encoder representations from transformers (BERT) were employed to design a toxic-fake news detection (FND) and classification system. As per the experiments, the linear SVM method outperformed BERT SVM, RF, and BERT RF with an accuracy of 92% and -score, -score, and -score of 95%, 85%, and 87%, respectively. Upon comparison, the proposed approach has either suppressed or achieved results very close to the state-of-the-art techniques in the literature by recording the best values on performance metrics such as accuracy, F1-score, precision, and recall for linear SVM. Overall, the proposed methods have shown promising results and urge further research to restrain toxic fake news. In contrast to prior research, the presented methodology leverages toxicity-oriented attributes and BERT-based sequence representations to discern toxic counterfeit news articles from nontoxic ones across social media platforms.

Designing a Testbed for Privacy-Preserving Voice-based Social Networks

Over the past decade, existing online social networks have emerged as vital platforms for fostering and nurturing social relationships. However, their reliance on text and multimedia-based interactions has posed limitations, particularly in terms of accessibility and hands-free operation, and in turn, impeded the full potential of online social networking. Voice-based social networks (VBSNs) offer a way to address these constraints by enabling multitasking through real-time audio communication, catering to a broader spectrum of users. To ensure the widespread adoption and practical utility of VBSNs, it is imperative that voice-based social networks and their applications offer secure and privacy-preserving services. In this paper, we outline the urgent requirements for voice-based social networks and evaluate existing open-source online social networking software and platforms to find the most suitable one for developing a testbed for VBSNs. Employing the chosen platform, we propose a design for VBSN testbed that facilitates the development and evaluation of VBSN applications while ensuring robust security and privacy measures through a blockchain-based decentralized identity and reputation mechanism. This proposed testbed serves as a valuable resource for researchers and developers to explore and address security and privacy concerns within the realm of voice-based social networking, enhancing the quality and reliability of their innovations.

How tobacco companies use the revolving door between government and industry to influence policymaking: an Australian case study.

Mixed-methods including non-experimental descriptive and exploratory case studies. To build a picture of tobacco lobbying through the revolving door in Australia, data was triangulated from multiple publicly available sources: 1) Australian federal, state and territory government lobbyist registers, 2) the online social networking platform, LinkedIn; and 3) Australian news media reports. Tobacco companies lobby the Australian government using 'in-house' employees, lobbyists working in firms acting on their behalf, and third-party allies with common interests. Almost half (48%) of internal tobacco company lobbyists had held positions in the Australian government (state, territory and/or federal) before or after working in the tobacco industry. Likewise, 55% of lobbyists acting on behalf of tobacco companies had held government positions before or after working as a lobbyist. In-house tobacco industry lobbyists, as well as those working on behalf of tobacco companies within lobbying firms, were found to have held senior governmental positions, such as a Member of Parliament (MP) or Senator, chief or deputy chief of staff, or senior advisor in a ministerial office, and many had moved into or out of government within 1 year of working for a tobacco company (56%) or as a lobbyist (48%). Tobacco companies are strategically using the revolving door between the government and the tobacco industry as a key political lobbying mechanism to try to influence public health policy in Australia.

CALEB: A Conditional Adversarial Learning Framework to enhance bot detection

The high growth of Online Social Networks (OSNs) over the last few years has allowed automated accounts, known as social bots, to gain ground. As highlighted by other researchers, many of these bots have malicious purposes and tend to mimic human behavior, posing high-level security threats on OSN platforms. Moreover, recent studies have shown that social bots evolve over time by reforming and reinventing unforeseen and sophisticated characteristics, making them capable of evading the current machine learning state-of-the-art bot detection systems. This work is motivated by the critical need to establish adaptive bot detection methods in order to proactively capture unseen evolved bots towards healthier OSNs interactions. In contrast with most earlier supervised ML approaches which are limited by the inability to effectively detect new types of bots, this paper proposes CALEB, a robust end-to-end proactive framework based on the Conditional Generative Adversarial Network (CGAN) and its extension, Auxiliary Classifier GAN (AC-GAN), to simulate bot evolution by creating realistic synthetic instances of different bot types. These simulated evolved bots augment existing bot datasets and therefore enhance the detection of emerging generations of bots before they even appear. Furthermore, we show that our augmentation approach overpasses other earlier augmentation techniques which fail at simulating evolving bots. Extensive experimentation on well established public bot datasets, show that our approach offers a performance boost of up to 10% regarding the detection of new unseen bots. Finally, the use of the AC-GAN Discriminator as a bot detector, has outperformed former ML approaches, showcasing the efficiency of our end to end framework.

Techniques for predicting dark web events focused on the delivery of illicit products and ordered crime

Malicious actors, specially trained professionals operating anonymously on the dark web (DW) platform to conduct cyber fraud, illegal drug supply, online kidnapping orders, CryptoLocker induction, contract hacking, terrorist recruitment portals on the online social network (OSN) platform, and financing are always a possibility in the hyperspace. The amount and variety of unlawful actions are increasing, which has prompted law enforcement (LE) agencies to develop efficient prevention tactics. In the current atmosphere of rapidly expanding cybercrime, conventional crime-solving methods are unable to produce results due to their slowness and inefficiency. The methods for accurately predicting crime before it happens "automated machine" to help police officers ease the burden on personnel while also assisting in preventing offense. To achieve and explain the results of a few cases in which such approaches were applied, we advise combining machine learning (ML) with computer vision (CV) strategies. This study's objective is to present dark web crime statistics and a forecasting model for generating alerts of illegal operations like drug supply, people smuggling, terrorist staffing and radicalization, and deceitful activities that are connected to gangs or organizations showing online presence using ML and CV to help law enforcement organizations identify, and accumulate proactive tactics for solving crimes.

AN INTEGRATED DEEP LEARNING APPROACH FOR SENTIMENT ANALYSIS ON TWITTER DATA

Analysis of Sentiment (SA) is a computational technique that seeks to extract subjective evaluations, attitudes, and emotional states from online platforms, specifically social media sites like Twitter. The subject has gained significant traction within the research community. The predominant emphasis of traditional sentiment analysis lies in the analysis of textual data. Twitter is widely recognized as a prominent online social networking platform that facilitates microblogging, wherein users share updates pertaining to various subjects through concise messages known as tweets. Twitter is a widely utilized platform that enables individuals to articulate their perspectives and emotions through the medium of tweets. Sentiment analysis refers to the computational methods of classification of given data in text format, categorizing it as either positive, negative, or neutral. The main goal of this study is to use deep learning methods for SA. The purpose is to guess sentiment and then evaluate the results based on accuracy, recall, and f-score. In this paper, a hybrid approach combining Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) techniques, specifically referred to as PSOGA, is proposed to optimize features for a modified neural network (MNN). The final step is to use the K-fold cross-validation method to assess the results. The dataset was obtained through the utilization of the Ruby Twitter API. The ultimate outcome is juxtaposed with the preceding Cuckoo Search (CS) algorithm that had been terminated.

DeepMUI: A novel method to identify malicious users on online social network platforms

SummaryThe use of online social network (OSN) platforms has become an essential component of contemporary society, facilitating global connectivity, and information sharing among individuals. The proliferation of malicious users has emerged as a noteworthy obstacle, exerting a detrimental effect on the authenticity of the data disseminated through these channels. A malicious profile is created with the intention of disseminating false information, manipulating perspectives, and executing harmful actions, including phishing schemes, identity theft, and the propagation of malware. Consequently, the identification of malicious users has emerged as an essential undertaking for both OSN platforms and researchers. The objective of this study is to investigate the issue of identifying malicious users on OSN platforms. The DeepMUI model has been introduced as a new approach to identifying malicious users on OSN platforms, utilizing user profile metadata‐derived characteristics. The DeepMUI architecture is composed of long short‐term memory and convolutional neural network models. Additionally, it integrates alterations to the pooling layer to improve its overall efficacy. The experiments have demonstrated that DeepMUI exhibits promising results in the task of identifying malicious users, with greater accuracy and minimal loss compared to existing methods.

Depolarization of opinions on social networks through random nudges.

Polarization of opinions has been empirically noted in many online social network platforms. Traditional models of opinion dynamics, based on statistical physics principles, do not account for the emergence of polarization and echo chambers in online network platforms. A recently introduced opinion dynamics model that incorporates the homophily factor-the tendency of agents to connect with those holding similar opinions as their own-captures polarization and echo chamber effects. In this work, we provide a nonintrusive framework for mildly nudging agents in an online community to form random connections. This is shown to lead to significant depolarization of opinions and decrease the echo chamber effects. Though a mild nudge effectively avoids polarization, overdoing this leads to another undesirable effect, namely, radicalization. Further, we obtain the optimal nudge probability to avoid the extremes of polarization and radicalization outcomes.