An Online Support Vector Machine Algorithm for Dynamic Social Network Monitoring

Characterizing user navigation and interactions in online social networks

ABSTRACTIn the last few decades sociologists were trying to explain human behaviour by analysing social networks, which requires access to data about interpersonal relationships. This represented a big obstacle in this research field until the emergence of online social networks (OSNs), which vastly facilitated the process of collecting such data. Nowadays, by crawling public profiles on OSNs, it is possible to build a social graph where “friends” on OSN become represented as connected nodes. OSN connection does not necessarily indicate a close real-life relationship, but using OSN interaction records may reveal real-life relationship intensities, a topic which inspired a number of recent researches. Still, published research currently lacks an extensive exploratory analysis of OSN interaction records, i.e. a comprehensive overview of users’ interaction via different ways of OSN interaction. In this paper, we provide such an overview by leveraging results of conducted extensive social experiment which managed to collect records for over 3200 Facebook users interacting with over 1,400,000 of their friends. Our exploratory analysis focuses on extracting population distributions and correlation parameters for 13 interaction parameters, providing valuable insight into OSN interaction for future researches aimed at this field of study.

This dissertation consists of two original studies in social networks and one original study in political economy. In the first two chapters, I study (i) how social networks form, and (ii) how economic agents optimize their behaviors for a given network structure. In the last chapter, I examine how election rules affect individual voting decisions and ultimate election outcomes. In Chapter 1, Social Network Formation and Strategic Interaction in Large Networks, I present a dynamic network formation model that aims to explain why some empirical degree distributions exhibit the increasing hazard rate property (IHRP). In my model, a sequentially arriving node forms a link with one existing node through a bilateral agreement. newborn node prefers a highly linked node; however, the more links an existing node has, the more the marginal return from an additional link diminishes. I prove that the IHRP emerges if and only if the latter effect prevails over the former. I present two implications of the IHRP for strategic interactions in networks. First, when there is uncertainty about neighboring agents' connectivity, the IHRP guarantees that a unique Bayesian equilibrium exists in a network game with strategic complementarities. Second, the IHRP characterizes a monotone revenue-maximizing mechanism with allocative externalities. In Chapter 2, Monopoly Pricing and Diffusion of a (Social) Network Good, I present a model of dynamic pricing and diffusion of a network good sold by a monopolist. In the model, the network good is a subscription social network good. This means that in each period, each consumer has to pay a subscription price to use the good, and the utility derived from subscribing to the good increases as more of her neighboring consumers subscribe. Consumers myopically optimize their subscription decisions, and the monopolist chooses a sequence of subscription prices that maximizes his discounted sum of per-period profits. Three main results emerge. First, I characterize a unique steady state of the monopoly market. Second, I find that optimal sequences of subscription prices oscillate around the subscription price at the steady state as time passes. Third, I analyze how changes in the monopolist's discount factor and the density of the social network affect the subscription price, subscription rate, and deadweight loss at the steady state. In Chapter 3, A Model of Pre-Electoral Coalition Formation, I study how two different election rules, simple plurality (e.g., as in South Korea) and two-round runoff (e.g., as in France), affect political candidates’ incentives to form pre-electoral coalitions (PECs). In my model, three candidates compete for a single office, and two candidates can form a PEC. Since the candidates are both policy- and office-motivated, one candidate can incentivize the other candidate to withdraw his candidacy by choosing a joint policy platform. I find that PECs are more likely to form in plurality elections than in two-round runoff elections. I further examine how other electoral environments, such as ideological distance and pre-election polls, influence incentives to form PECs.

Online social networks (forum) have become extremely popular; numerous sites allow users to interact and share content using social links observing that there are some limitations on most of the social interaction networks. A recent study has demonstrated that the “strength of ties” varies widely, ranging from pairs of users who are best friends to pairs of users who even wished they weren’t friends. In order to distinguish between these strong and weak links, researchers have suggested examining the activity networks, the network that is formed by users who actually interact using one or many of the methods provided by the social networking site. Also for effective study, we collect data on both friendship relationship and interactions for a large subset of the Facebook social network and forum, Twitter, 2go, m.eskimi and other social interactive networks. From a technological standpoint, forums or boards are web applications managing user-generated content. Early Internet forums could be described as a web version of an electronic mailing list or newsgroup (such as exist on Usenet); allowing people to post messages and comment on other messages. It has been discovered that Chat forum is one of the newest way in which organization and the business world use now to reach out to their numerous customers in order to get feedback and responses concerning its entire product lines to share of their opinion.

Huberized one-class support vector machine with truncated loss function in the primal space

With the increasing development of online social networks, privacy protection attracts more and more concerns. Most research on privacy protection focus on specific techniques, and they assume users adopt security protection mechanisms for granted. The impact of user behaviors on the deployment of these mechanisms is usually neglected. In fact, the actual effect of privacy protection mechanisms depends heavily on user behaviors. Whether a user adopts the security mechanism is affected by incentives, interactions with friends and other factors. It is necessary to encourage users to take privacy protection settings in order to improve the security environment in social networks. Therefore, in this paper, a network evolutionary game model is proposed for studying the influence of user behaviors on the deployment of privacy protection mechanisms, where whether to take privacy settings is used as game strategies. Considering real interactions in social networks, games only happen between connected users in the model. Then the dynamic equation of user behaviors over time is derived. Based on the dynamic equation, evolutionary stability conditions are proven according to the benefit cost ratio of taking the protection. Experiments in real social networks verify the feasibility of the proposed model. The analysis results are helpful for guiding security services to promote the deployment of privacy protection mechanisms in social networks.

The traditional template matching strategy of optical and synthetic aperture radar (SAR) is sensitive to the nonlinear transformation between two images. In some cases, the optical and SAR image pairs do not conform to the affine transformation condition. To address this issue, this study presents a novel template matching strategy which uses the One-Class Support Vector Machine (SVM) to remove outliers. First, we propose a method to construct the similarity map dataset using the SEN1-2 dataset for training the One-Class SVM. Second, a four-step strategy for optical and SAR image registration is presented in this paper. In the first step, the optical image is divided into some grids. In the second step, the strongest Harris response point is selected as the feature point in each grid. In the third step, we use Gaussian pyramid features of oriented gradients (GPOG) descriptor to calculate the similarity map in the search region. The trained One-Class SVM is used to remove outliers through similarity maps in the fourth step. Furthermore, the number of improve matches (NIM) and the rate of improve matches (RIM) are designed to measure the effect of image registration. Finally, this paper designs two experiments to prove that the proposed strategy can correctly select the matching points through similarity maps. The experimental results of the One-Class SVM in dataset show that the One-Class SVM can select the correct points in different datasets. The image registration results obtained on the second experiment show that the proposed strategy is robust to the nonlinear transformation between optical and SAR images.

Popular online social networks (OSNs) like Facebook and Twitter are changing the way users communicate and interact with the Internet. A deep understanding of user interactions in OSNs can provide important insights into questions of human social behavior, and into the design of social platforms and applications. However, recent studies have shown that a majority of user interactions on OSNs are latent interactions, passive actions such as profile browsing that cannot be observed by traditional measurement techniques. In this paper, we seek a deeper understanding of both visible and latent user interactions in OSNs. For quantifiable data on latent user interactions, we perform a detailed measurement study on Renren, the largest OSN in China with more than 150 million users to date. All friendship links in Renren are public, allowing us to exhaustively crawl a connected graph component of 42 million users and 1.66 billion social links in 2009. Renren also keeps detailed visitor logs for each user profile, and counters for each photo and diary/blog entry. We capture detailed histories of profile visits over a period of 90 days for more than 61,000 users in the Peking University Renren network, and use statistics of profile visits to study issues of user profile popularity, reciprocity of profile visits, and the impact of content updates on user popularity. We find that latent interactions are much more prevalent and frequent than visible events, non-reciprocal in nature, and that profile popularity are uncorrelated with the frequency of content updates. Finally, we construct latent interaction graphs as models of user browsing behavior, and compare their structural properties against those of both visible interaction graphs and social graphs.

Online Social Network (OSN) is like a virtual community where people build social networks and relations with one another. The open access to the Internet has increased the growth of OSN which has attracted intruders to exploit the weaknesses of the Internet and OSN to their own gain. The rise in the usage of OSN has posed security threats to OSN users as they share personal and sensitive information online which could be exploited by these intruders by creating profiles to carry out a series of malicious activities on the social network. In fact, it is no gain saying that the intent of creating fake accounts has adverse effect and the Internet has made it quite easy to concede one’s identity; and this makes it difficult to detect fake accounts as they try to imitate real accounts. In this study, a model that can accurately identify fake profiles in OSN which uses Natural Language Processing Technique to eliminate or reduce the size of the dataset thereby improving the overall performance of the model was proposed. Principal Component Analysis was used for appropriate feature selection. After extraction, six attributes/features that influenced the classifier were found. Support Vector Machine (SVM), Naïve Bayes and Improved Support Vector Machine (ISVM) were used as Classifiers. ISVM introduced a penalty parameter to the standard SVM objective function to reduce the inequality constraints between the slack variables. This gave a better result of 90% than the SVM and Naïve Bayes which gave 77.4% and 77.3% respectively.

The main objective of structural health monitoring is to provide reliable information about the health state of the critical structures by implementing a damage characterization strategy to detect the presence of damage, location, severity as possibly failure prediction as soon as the damage occurs. This paper presents a robust approach to detect and characterize a gradually evolving damage based on time responses data captured from a steel reinforced concrete structure. The presented method is in the context of unsupervised and non-model-based approaches, hence, there is no need for any representative numerical/finite element model of the structure to be built. In this work, we propose one-class support vector machine as an anomaly detection method. One-class support vector machine fits well for damage diagnosis in structural health monitoring since there may exist many damaged patterns and one-class support vector machine can detect all of them as anomalies. To demonstrate the feasibility of the method in the detection and assessment of a gradually evolving deterioration, a test bed was established to replicate a concrete jack arch which is a main structural component on the Sydney Harbour Bridge – one of Australia’s iconic structures, see Figure 1. The structure is a concrete cantilever beam with an arch section which is located on the eastern side of the bridge underneath the bus lane. It is assumed that the structure is subjected to Gaussian white noise excitation. The response of the structure was collected by 10 uniaxial accelerometers placed in frontal face of the jack arch. The measurement was conducted for 2 seconds at a sampling rate of 8 kHz, resulting in 16000 samples for each event. Figure 1 Illustration of the test bed. https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781315207681/cd556cd4-4dcf-4efe-8e29-56fc67b8bfbd/content/fig321_1.jpg/> A crack is introduced in the structure using a cutting saw and its length is progressively increased in four stages while the depth was constant; these four damage cases correspond to less than 0.5% reduction in the first three modes of the structure. Figure 2 presents the average of all decision values obtained from all sensors for healthy state and different damage states using one-class support vector machine. As seen, the method is able to successfully separate the healthy state from the damage states and as the severity of damage increases, higher negative decision values are obtained. Figure 2 Damage detection using decision values of all test events. https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781315207681/cd556cd4-4dcf-4efe-8e29-56fc67b8bfbd/content/fig321_2.tif/>

In this paper, we propose a method by engaging the One Class Support Vector Machine (OC-SVM) in the identification of Diffractive Optically Variable Images (DOVIs). OC-SVM, as a special SVM, can solve the problems of high-dimensional data sets and small sample size (SSS) with positive and negative unbalance training data. Image feature matrix is built by extracting image features from texture aspects. OC-SVM can be trained with the high-dimensional matrix directly, and does not have to reduce the dimensionality of feature matrix as the usual methods. The experiment results show the effectiveness of the proposed approach against Linear Discriminant Analysis. Considering time cost and correct classification rate, OC-SVM is suitable for the identification of DOVIs.

One-class support vector machine (OCSVM) is one of the most popular algorithms in the one-class classification problem, but it has one obvious disadvantage: it is sensitive to noise. In order to solve this problem, the fuzzy membership degree is introduced into OCSVM, which makes the samples with different importance have different influences on the determination of classification hyperplane and enhances the robustness. In this paper, a new calculation method of membership degree is proposed and introduced into the fuzzy multiple kernel OCSVM (FMKOCSVM). The combined kernel is used to measure the local similarity between samples, and then, the importance of samples is determined based on the local similarity between training samples, so as to determine the membership degree and reduce the impact of noise. The proposed membership requires only positive data in the calculation process, which is consistent with the training set of OCSVM. In this method, the noise has a smaller membership value, which can reduce the negative impact of noise on the classification boundary. Simultaneously, this method of calculating membership has a higher efficiency. The experimental results show that FMKOCSVM based on proposed local similarity membership is efficient and more robust to outliers than the ordinary multiple kernel OCSVMs.

In online social networks, numerous users generate and distribute tremendous content simultaneously, and how information spreads in online society has attracted critical attention. Recent studies introduce deep learning to help mine the complex diffusion pattern and forecast the diffusion trend. However, there are two limitations of the state-of-art methods. On the one hand, existing models only focus on the internal influence within the diffusion flow but ignore the external influence from the dissemination of other contents. On the other hand, the dynamics of user interest are barely considered, while the changes of user preference for contents also have a considerable impact on future diffusion. To address these issues, we introduce the hypergraph structure and a sequential framework to model complex interactions in social networks. Then, we propose dual-channel hypergraph neural networks to tackle the diffusion prediction problem, denoted by HyperINF. Specifically, in the user channel, we build sequential user interactive hypergraphs and learn the dynamic user representation, and in the diffusion channel, we construct a diffusion interactive graph to capture the cross-diffusion relation. At last, we consider the social relation to help make the prediction. Experimental results on three datasets suggest the effectiveness and practicability of the proposed framework.

Languaging when contexts collapse: Audience design in social networking

Soft sensors are widely used to estimate process variables that are difficult to measure online. By using soft sensors, analyzer faults can be detected by estimation errors. However, it is difficult to detect abnormal data and determine the reasons because estimation errors increase not only due to analyzer faults but also due to variations caused by changes in the state of chemical plants. To separate those factors, we previously proposed to construct the relationships between distances to soft sensor models (DMs) and the accuracy of prediction of the models quantitatively and estimate the prediction accuracy of new data online. In this article, we used a one‐class support vector machine (OCSVM) to estimate data density and the output of an OCSVM as a DM. The proposed method was applied to real industrial data and the superiority of the proposed DM to the traditional ones was demonstrated by comparing their results. © 2013 American Institute of Chemical Engineers AIChE J, 59: 2046–2050, 2013