Data Summaries and Representations: Definitions and Practical Use

Community detection is an effective approach to unveil relationships among individuals in online social networks. In the literature, quite a few algorithms have been proposed to conduct community detection by exploiting the topology of social networks and the attributes of social actors. In practice, community detection is usually conducted by third parties, such as advertisement companies and hospitals, with access to social networks for different purposes, which can easily lead to a privacy breach. In this paper, we investigate community detection in social networks aiming to protect the privacy of both the network topology and the users' attributes. We show that with additional prior knowledge, community detection can be performed by querying the information of only a fraction of instead of the entire population. In particular, we first propose a new scheme called differentially private community detection (DPCD). DPCD detects communities in social networks via a probabilistic generative model, which can be decomposed into subproblems solved by individual users. The private social relationships and attributes of each user are protected by objective perturbation with differential privacy guarantees. Then, we propose a parsimonious node affiliation recovery (NAR) algorithm, which is also differentially private, to unveil the community affiliation information of the whole population based on that of the limited number of queried individuals by solving a sparse optimization problem. Through both theoretical analysis and experimental validation using synthetic and real-world social networks, we demonstrate that the proposed DPCD scheme detects social communities under the modest privacy budget. In addition, we show the effectiveness of NAR to perform community detection by querying a limited number of individuals in social networks.

Identifying the k most individuals in a social network is a well-studied problem. The objective is to detect k individuals in a (social) network who will influence the maximum number of people, if they are independently convinced of adopting a new strategy (product, idea, etc). There are cases in real life, however, where we aim to instigate instead of individuals to trigger network diffusion. Such cases abound, e.g., billboards, TV commercials and newspaper ads are utilized extensively to boost the popularity and raise awareness. In this paper, we generalize the influential nodes problem. Namely we are interested to locate the most influential groups in a network. As the first paper to address this problem: we (1) propose a fine-grained model of information diffusion for the group-based problem, (2) show that the process is submodular and present an algorithm to determine the under this model (with a precise approximation bound), (3) propose a coarse-grained model that inspects the network at group level (not individuals) significantly speeding up calculations for large networks, (4) show that the diffusion function we design here is submodular in general case, and propose an approximation algorithm for this coarse-grained model, and finally by conducting experiments on real datasets, (5) demonstrate that seeding members of selected to be the first adopters can broaden diffusion (when compared to the individuals case). Moreover, we can identify these much faster (up to 12 million times speedup), delivering a practical solution to this problem.

Many patterns have been uncovered in complex systems through the application of concepts and methodologies of complex networks. Unfortunately, the validity and accuracy of the unveiled patterns are strongly dependent on the amount of unavoidable noise pervading the data, such as the presence of homonymous individuals in social networks. In the current paper, we investigate the problem of name disambiguation in collaborative networks, a task that plays a fundamental role on a myriad of scientific contexts. In special, we use an unsupervised technique which relies on a particle competition mechanism in a networked environment to detect the clusters. It has been shown that, in this kind of environment, the learning process can be improved because the network representation of data can capture topological features of the input data set. Specifically, in the proposed disambiguating model, a set of particles is randomly spawned into the nodes constituting the network. As time progresses, the particles employ a movement strategy composed of a probabilistic convex mixture of random and preferential walking policies. In the former, the walking rule exclusively depends on the topology of the network and is responsible for the exploratory behavior of the particles. In the latter, the walking rule depends both on the topology and the domination levels that the particles impose on the neighboring nodes. This type of behavior compels the particles to perform a defensive strategy, because it will force them to revisit nodes that are already dominated by them, rather than exploring rival territories. Computer simulations conducted on the networks extracted from the arXiv repository of preprint papers and also from other databases reveal the effectiveness of the model, which turned out to be more accurate than traditional clustering methods.

The objective of this thesis is to promote the employment of Linked Open Data practices in the creation and presentation of digital editions of written historical sources, to offer tools for facilitating research, and to suggest a new research method based on the Linked Data representation of digital editions. Even though the current thesis mainly focuses on proposing a digital approach to research in the field of history of science, the discussed concepts and tools, and the new method are applicable to any humanities field. 
\n
\nIn order to understand the needs of the field of history of science and to develop tools with practical use in research, I chose to study the contributions of Jacob Bernoulli (1654–1705) to rational mechanics, a topic that has not been previously studied. The first part of this thesis is dedicated to the study of Bernoulli’s works on mechanics mainly focusing on his analyses of the Funicularia problem, elasticity, theory of neutral axis, and parallelogram of forces. This research is based on an in-depth study of Bernoulli’s mechanics through published articles, correspondences, and private notes.
\n 
\nThe second part of this thesis describes the process of creating digital editions of these types of historical data using semantic web technologies, their integration into a virtual research environment (VRE) called the Bernoulli Euler Online (BEOL) platform, and the tools offered to study the generated network of resources. This thesis illustrates the representation of the scientific-historical data as a network of resources on the web and describes the advanced research tools that allow for powerful analysis of the RDF data graphs. Since natural philosophers of the early modern period communicated ideas through correspondence creating what is known today as the republic of letters, the digital edition of their correspondence available on different online portals must also be connected to create a virtual republic of letters. This thesis explains a generic e-infrastructure that I have developed to connect the digital editions of other natural philosophers’ correspondence presented on external platforms, such as The Newton Project and the Briefportal Leibniz, to the BEOL platform without centralizing data silos. This infrastructure enables researchers to access and analyze the entire correspondence network through one central platform. This thesis thus suggests a new research method for the history of the science field — the network method — that is based on the network representation of research data. Lastly, this thesis suggests a new technique to visualize RDF-based data as web-based interactive 3D graphs that can be used for any humanities data. The purpose of this tool is to strengthen the suggested network method by giving a real picture of the relational structure of data and allowing researchers to interact with a 3D model of the network resources.

Facilities provided by social network sitesallow individuals to use social networks for various purposes such ascommunication, entertainment and information sharing. These purposes varydepending on the social networks individuals use and on their personalitytraits. The present study examined preservice teachers’ social network usepurposes with respect to their personality traits. The participants of thestudy were 660 preservice teachers attending a Turkish university. For datacollection in the study “Scale of Social Network Use Purposes” and “BasicPersonality Traits Inventory” were used. The research data were analyzed using PearsonCorrelation Analysis and Multi-Linear Regression Analysis. The findingsrevealed that among the personality traits, conscientiousness, openness to experienceand negative valence were predictors of the sub-factors of social network usepurposes. As for other personality factors, neuroticism and extraversion werenot significant predictors of the total social network use purpose score or itssub-factors.

The fast development of artificial intelligence (AI) and machine learning (ML) have made it possible to make substantial advances in automated skin-cancer diagnosis, but its use in practice is still scarce. This paper focuses on three unresolved issues in AI-assisted dermatology by systematically reviewing recent progress in the field, such as data diversity, model interpretability, and clinical integration. The articles reviewed in 20172024 were analyzed according to a systematic search of medical-imaging databases and IEEE/Scopus repositories. The research shows that the main gaps that require attention include the lack of balance of the datasets within the demographics of skin-tone and the obscurity of deep-learning models, as well as regulatory or workflow obstacles as impediments to clinical validation. Thematic synthesis and comparative information about state-of-the-art methods, interpretability techniques, and fairness frameworks are described. Lastly, the review also suggests a roadmap for responsible, explainable, and workflow-centric AI systems in dermatology. The results indicate that success in the future lies not only in the accuracy of the model but also in fairness in the representation of data, clear decision-making, and flawless clinical usability.

The Internet and the World Wide Web have become two key components in today's technology based organizations and businesses. As the Internet is becoming more and more popular, it is starting to make a big impact on people's day-to-day life. As a result of this revolutionary transformation towards the modern technology, social networking on the World Wide Web has become an integral part of a large number of people's lives. Social networks are websites which allow users to communicate, share knowledge about similar interests, discuss favorite topics, review and rate products/services, etc. These websites have become a powerful source in shaping public opinion on virtually every aspect of commerce. Marketers are challenged with identifying influential individuals in social networks and connecting with them in ways that encourage viral marketing content movement and there has been little empirical research study about of this website to diffuse of viral marketing content. In this article, we explore the role of social network websites which has influence on viral marketing, and the characteristics of the most influential users to spread share viral content. Structural equation modeling is used to examine the patterns of inter-correlations among the constructions and to empirically test the hypotheses.

Identifying authoritative individuals is a well-known approach in extracting actionable knowledge, known as “Knowledge Representation”, in a social network. Previous researches suggest measures to identify influential individuals, however, such individuals might not represent the appropriate context (relationships, interactions, etc.). For example, it is nearly an impossible task for a single individual to organize a mass protest of the scale of Occupy Wall Street. Similarly, other events such as the Arab Spring, coordinating crisis responses for natural disasters (e.g., the Haiti earthquake), or even organizing flash mobs would require a key set of individuals rather than a single or the most authoritative one. These events demonstrate the need and importance of examining influential structures rather than single individuals in social networks. A new methodology is proposed to identify such influential structures and recognizing their importance. The proposed methodology is evaluated empirically with real-world data from NIST's Tweets2011 corpus. We also introduce a novel and objective evaluation strategy to ascertain the efficacy of the focal patterns. Challenges with future research directions are outlined.

Communities in social networks form with different purposes and play a significant role in interpersonal interactions. Analysis of virtual communities indicates a more precise understanding of the behaviours and desires of individuals in social networks. In this paper, new measures have been proposed for analyzing implicit and explicit communities in Online Social Networks (OSNs). The measures of “potential value of the community members” and “value of the community messages”, which are used for calculating the measure of “community value” are among the most important measures introduced in this paper. Another measure introduced is “user influence rate” in a community, which represents the contribution of a person in creating value in a community. To provide a sound dataset, we collected the information from several real implicit communities in Twitter based on different hashtags. Finally, the suggested measures have been analyzed and compared statistically and behaviourally across different communities. The results of this research well indicate the importance and practicality of the measures introduced in Community analysis of Twitter.

We focus on analysis of dominance, power, influence - that by definition asymmetric - between pairs of individuals in social networks. We conduct dominance analysis based on the canonical analysis of asymmetry that decomposes a square asymmetric matrix into two parts, a symmetric one and a skew-symmetric one, and then applies the singular value decomposition (SVD) on the skew-symmetric part. Each individual node can be projected as one 2-dimensional point based on its row values at each pair of successive singular vectors. The asymmetric relationship between two individuals can then be captured by areas of triangles formed from the two points and the origin in each 2-dimensional space. We quantify node dominance (submissive) score based on the relative position of the node's coordinate from coordinates of all other nodes it dominates (subdues) in the projected singular vector spaces. We conduct dominance/submissiveness analysis for several representative networks including perfect linear orderings, networks with tree structure, and networks with random graphs and examine the departures of a real social network from those representative graphs. Empirical evaluations demonstrate the effectiveness of the proposed approach.

WITHDRAWN: Experimental investigation of automated system for twitter sentiment analysis to predict the public emotions using machine learning algorithms

A new algorithm for positive influence maximization in signed networks

Individuals in social networks are often organized under some hierarchy such as a command structure. In many cases, when this structure is unknown, there is a need to discover hierarchical organization. In this paper, we propose a novel, simple, and flexible method based on maximum likelihood to infer social hierarchy from weighted social networks. We empirically evaluate our method against both simulated and real-world datasets and show that our approach accurately recovers the underlying, latent hierarchy.

Identifying individuals associated with organized criminal networks: A social network analysis

Math Education for America? analyzes math education policy through the social network of individuals and private and public organizations that influence it in the United States. The effort to standardize a national mathematics curriculum for public schools in the U.S. culminated in 2010 when over 40 states adopted the Common Core State Standards for Mathematics. Rather than looking at the text of specific policy documents, this book complements existing critical reviews of the national math education curriculum by employing a unique social network analysis. Breaking new ground in detailing and theorizing the politics of math education, Wolfmeyer argues that the private interests of this network are closely tied to a web of interrelated developments: human capital education policy, debates over traditional and reform pedagogy, the assumed content knowledge deficit of math teachers, and the proliferation of profit-driven educational businesses. By establishing the interconnectedness of these interests with the national math education curriculum, he shows how the purported goals of math education reform are aligned with the prevailing political agendas of this social network rather than the national interest.