Modelli e linguaggi per grafi multi-livello

I grafi sono molto utilizzati per la rappresentazione di dati, sopratutto in quelle aree dove l’informazione sull’interconnettività e la topologia dei dati è importante tanto quanto i dati stessi, se non addirittura di più. Ogni area di applicazione ha delle proprie necessità, sia in termini del modello che rappresenta i dati, sia in termini del linguaggio capace di fornire la necessaria espressività per poter fare interrogazione e trasformazione dei dati. È sempre più frequente che si richieda di analizzare dati provenienti da diversi sistemi, oppure che si richieda di analizzare caratteristiche dello stesso sistema osservandolo a granularità differenti, in tempi differenti oppure considerando relazioni differenti. Il nostro scopo è stato quindi quello di creare un modello, che riesca a rappresentare in maniera semplice ed efficace i dati, in tutte queste situazioni. Entrando più nei dettagli, il modello permette non solo di analizzare la singola rete, ma di analizzare più reti, relazionandole tra loro. Il nostro scopo si è anche esteso nel definire un’algebra, che, tramite ai suoi operatori, permette di compiere delle interrogazioni su questo modello. La definizione del modello e degli operatori sono stati maggiormente guidati dal caso di studio dei social network, non tralasciando comunque di rimanere generali per fare altri tipi di analisi. In seguito abbiamo approfondito lo studio degli operatori, individuando delle proprietà utili per fare delle ottimizzazioni, ragionando sui dettagli implementativi, e fornendo degli algoritmi di alto livello. Per rendere più concreta la definizione del modello e degli operatori, in modo da non lasciare spazio ad ambiguità, è stata fatta anche un’implementazione, e in questo elaborato ne forniremo la descrizione.

On the k-anonymization of time-varying and multi-layer social graphs

The popularity of online social media platforms provides an unprecedented opportunity to study real-world complex networks of interactions. However, releasing this data to researchers and the public comes at the cost of potentially exposing private and sensitive user information. It has been shown that a naive anonymization of a network by removing the identity of the nodes is not sufficient to preserve users’ privacy. In order to deal with malicious attacks, k -anonymity solutions have been proposed to partially obfuscate topological information that can be used to infer nodes’ identity. In this paper, we study the problem of ensuring k anonymity in time-varying graphs, i.e., graphs with a structure that changes over time, and multi-layer graphs, i.e., graphs with multiple types of links. More specifically, we examine the case in which the attacker has access to the degree of the nodes. The goal is to generate a new graph where, given the degree of a node in each (temporal) layer of the graph, such a node remains indistinguishable from other k-1 nodes in the graph. In order to achieve this, we find the optimal partitioning of the graph nodes such that the cost of anonymizing the degree information within each group is minimum. We show that this reduces to a special case of a Generalized Assignment Problem, and we propose a simple yet effective algorithm to solve it. Finally, we introduce an iterated linear programming approach to enforce the realizability of the anonymized degree sequences. The efficacy of the method is assessed through an extensive set of experiments on synthetic and real-world graphs.

Clustering algorithms for anti-money laundering using graph theory and social network analysis

HEMOLIA (a project under European community’s 7th framework programme) is a new generation Anti-Money Laundering (AML) intelligent multi-agent alert and investigation system which in addition to the traditional financial data makes extensive use of modern society’s huge telecom data source, thereby opening up a new dimension of capabilities to all Money Laundering fighters (FIUs, LEAs) and Financial Institutes (Banks, Insurance Companies, etc.). This Master-Thesis project is done at AIA, one of the partners for the HEMOLIA project in Barcelona. The objective of this thesis is to find the clusters in a network drawn by using the financial data. An extensive literature survey has been carried out and several standard algorithms related to networks have been studied and implemented. The clustering problem is a NP-hard problem and several algorithms like K-Means and Hierarchical clustering are being implemented for studying several problems relating to sociology, evolution, anthropology etc. However, these algorithms have certain drawbacks which make them very difficult to implement. The thesis suggests (a) a possible improvement to the K-Means algorithm, (b) a novel approach to the clustering problem using the Genetic Algorithms and (c) a new algorithm for finding the cluster of a node using the Genetic Algorithm.

Understanding Online Game Players' Post-adoption Behavior: an Investigation of Social Network Games in China

Retaining players and re-attracting switching players has long been a central topic for SNG providers with regard to the post-adoption stage of playing an online game. However, there has not been much research which has explored players’ post-adoption behavior by incorporating the continuance intention and the switching intention. In addition, traditional IS continuance theories were mainly developed to investigate users’ continued use of utilitarian IS, and thus they may fall short when trying to explain the continued use of hedonic IS. Furthermore, compared to the richer literature on IS continuance, far too little attention has been paid to IS switching, leading to a dearth of knowledge on the subject, despite the increased incidence of the switching phenomenon in the IS field. By addressing the limitations of prior literature, this study seeks to examine the determinants of SNG players’ two different post-adoption behaviors, including the continuance intention and the switching intention. This study takes a positivist approach and uses survey research method to test five proposed research models based on Unified Theory of Acceptance and Use of Technology 2; Use and Gratification Theory; Push-Pull-Mooring model; Cognitive Dissonance Theory; and a self-developed model respectively with empirical data collected from the SNG players of one of the biggest SNG providers in China. A total of 3919 valid responses and 541 valid responses were used to examine the continuance intention and the switching intention, respectively. SEM is utilized as the data analysis method. The proposed research models are supported by the empirical data. The continuance intention is determined by enjoyment, fantasy, escapism, social interaction, social presence, social influence, achievement and habit. The switching intention is determined by enjoyment, satisfaction, subjective norms, descriptive norms, alternative attractiveness, the need for variety, change experience, and adaptation cost. This study contributes to IS theories in three important ways. Firstly, it shows IS switching should be included in IS post-adoption research together with IS continuance. Secondly, a modern IS is usually multi-functional and SNG players have multiple reasons for using a SNG, thus a player’s beliefs about the hedonic, social and utilitarian perceptions of their continued use of the SNG exert significant effects on the continuance intention. Thirdly, the determinants of the switch ing intention mainly exert push, pull, and mooring effects. Players’ beliefs about their current SNG and the available alternatives, as well as their individual characteristics are all significant determinants of the switching intention. SNG players combine these effects in order to formulate the switching intention. Finally, this study presents limitations and suggestions for future research.

Community Detection in Multi-Layer Networks

In the scope of the current thesis we review and analyse networks that are formed by nodes with several attributes. We suppose that different layers of communities are embedded in such networks, besides each of the layers is connected with nodes' attributes. For example, examine one of a variety of online social networks: an user participates in a plurality of different groups/communities – schoolfellows, colleagues, clients, etc. We introduce a detection algorithm for the above-mentioned communities. Normally the result of the detection is the community supplemented just by the most dominant attribute, disregarding others. We propose an algorithm that bypasses dominant communities and detects communities which are formed by other nodes' attributes. We also review formation models of the attributed networks and present a Human Communication Network (HCN) model. We introduce a High School Texting Network (HSTN) and examine our methods for that network.

Stakeholder analysis combined with social network analysis provides fine-grained insights into water infrastructure planning processes

Environmental policy and decision-making are characterized by complex interactions between different actors and sectors. As a rule, a stakeholder analysis is performed to understand those involved, but it has been criticized for lacking quality and consistency. This lack is remedied here by a formal social network analysis that investigates collaborative and multi-level governance settings in a rigorous way. We examine the added value of combining both elements. Our case study examines infrastructure planning in the Swiss water sector. Water supply and wastewater infrastructures are planned far into the future, usually on the basis of projections of past boundary conditions. They affect many actors, including the population, and are expensive. In view of increasing future dynamics and climate change, a more participatory and long-term planning approach is required. Our specific aims are to investigate fragmentation in water infrastructure planning, to understand how actors from different decision levels and sectors are represented, and which interests they follow. We conducted 27 semi-structured interviews with local stakeholders, but also cantonal and national actors. The network analysis confirmed our hypothesis of strong fragmentation: we found little collaboration between the water supply and wastewater sector (confirming horizontal fragmentation), and few ties between local, cantonal, and national actors (confirming vertical fragmentation). Infrastructure planning is clearly dominated by engineers and local authorities. Little importance is placed on longer-term strategic objectives and integrated catchment planning, but this was perceived as more important in a second analysis going beyond typical questions of stakeholder analysis. We conclude that linking a stakeholder analysis, comprising rarely asked questions, with a rigorous social network analysis is very fruitful and generates complementary results. This combination gave us deeper insight into the socio-political-engineering world of water infrastructure planning that is of vital importance to our well-being.

The Relationship between Adolescent Social Network Structure and Perceptions of School Climate

Thesis (Ph.D.)--University of Washington, 2016-06

Mining the online social network data: Influence, summarization, and organization

Online Social Network (OSN) services provided by Internet companies bring people together to chat, share the information, and enjoy the information. Meanwhile, huge amounts of data are generated by those services (they can be regarded as the social media ) every day, every hour, even every minute, and every second. Currently, researchers are interested in analyzing the OSN data, extracting interesting patterns from it, and applying those patterns to real-world applications. However, due to the large-scale property of the OSN data, it is difficult to effectively analyze it. This dissertation focuses on applying data mining and information retrieval techniques to mine two key components in the social media data — users and user-generated contents. Specifically, it aims at addressing three problems related to the social media users and contents: (1) how does one organize the users and the contents? (2) how does one summarize the textual contents so that users do not have to go over every post to capture the general idea? (3) how does one identify the influential users in the social media to benefit other applications, e.g., Marketing Campaign? The contribution of this dissertation is briefly summarized as follows. (1) It provides a comprehensive and versatile data mining framework to analyze the users and user-generated contents from the social media. (2) It designs a hierarchical co-clustering algorithm to organize the users and contents. (3) It proposes multi-document summarization methods to extract core information from the social network contents. (4) It introduces three important dimensions of social influence, and a dynamic influence model for identifying influential users.

Mining the Online Social Network Data: Influence, Summarization, and Organization

Online Social Network (OSN) services provided by Internet companies bring people together to chat, share the information, and enjoy the information. Meanwhile, huge amounts of data are generated by those services (they can be regarded as the social media ) every day, every hour, even every minute, and every second. Currently, researchers are interested in analyzing the OSN data, extracting interesting patterns from it, and applying those patterns to real-world applications. However, due to the large-scale property of the OSN data, it is difficult to effectively analyze it. This dissertation focuses on applying data mining and information retrieval techniques to mine two key components in the social media data — users and user-generated contents. Specifically, it aims at addressing three problems related to the social media users and contents: (1) how does one organize the users and the contents? (2) how does one summarize the textual contents so that users do not have to go over every post to capture the general idea? (3) how does one identify the influential users in the social media to benefit other applications, e.g., Marketing Campaign? The contribution of this dissertation is briefly summarized as follows. (1) It provides a comprehensive and versatile data mining framework to analyze the users and user-generated contents from the social media. (2) It designs a hierarchical co-clustering algorithm to organize the users and contents. (3) It proposes multi-document summarization methods to extract core information from the social network contents. (4) It introduces three important dimensions of social influence, and a dynamic influence model for identifying influential users.

Empirical evidence for ultrametric structure in multi-layer perceptron error surfaces

Combinatorial optimization problems share an interesting property with spin glass systems in that their state spaces can exhibit ultrametric structure. We use sampling methods to analyse the error surfaces of feedforward multi-layer perceptron neural networks learning encoder problems. The third order statistics of these points of attraction are examined and found to be arranged in a highly ultrametric way. This is a unique result for a finite, continuous parameter space. The implications of this result are discussed.

Partial disjoint path for multi-layer protection in GMPLS networks

In this paper, different recovery methods applied at different network layers and time scales are used in order to enhance the network reliability. Each layer deploys its own fault management methods. However, current recovery methods are applied to only a specific layer. New protection schemes, based on the proposed partial disjoint path algorithm, are defined in order to avoid protection duplications in a multi-layer scenario. The new protection schemes also encompass shared segment backup computation and shared risk link group identification. A complete set of experiments proves the efficiency of the proposed methods in relation with previous ones, in terms of resources used to protect the network, the failure recovery time and the request rejection ratio

Enhanced multi-layer protection in multi-service GMPLS networks

This paper focuses on QoS routing with protection in an MPLS network over an optical layer. In this multi-layer scenario each layer deploys its own fault management methods. A partially protected optical layer is proposed and the rest of the network is protected at the MPLS layer. New protection schemes that avoid protection duplications are proposed. Moreover, this paper also introduces a new traffic classification based on the level of reliability. The failure impact is evaluated in terms of recovery time depending on the traffic class. The proposed schemes also include a novel variation of minimum interference routing and shared segment backup computation. A complete set of experiments proves that the proposed schemes are more efficient as compared to the previous ones, in terms of resources used to protect the network, failure impact and the request rejection ratio

Non-Linear Unsteady Aerodynamic Response Approximation Using Multi-Layer Functionals

Non-linear functional representation of the aerodynamic response provides a convenient mathematical model for motion-induced unsteady transonic aerodynamic loads response, that accounts for both complex non-linearities and time-history effects. A recent development, based on functional approximation theory, has established a novel functional form; namely, the multi-layer functional. For a large class of non-linear dynamic systems, such multi-layer functional representations can be realised via finite impulse response (FIR) neural networks. Identification of an appropriate FIR neural network model is facilitated by means of a supervised training process in which a limited sample of system input-output data sets is presented to the temporal neural network. The present work describes a procedure for the systematic identification of parameterised neural network models of motion-induced unsteady transonic aerodynamic loads response. The training process is based on a conventional genetic algorithm to optimise the network architecture, combined with a simplified random search algorithm to update weight and bias values. Application of the scheme to representative transonic aerodynamic loads response data for a bidimensional airfoil executing finite-amplitude motion in transonic flow is used to demonstrate the feasibility of the approach. The approach is shown to furnish a satisfactory generalisation property to different motion histories over a range of Mach numbers in the transonic regime.

Partial disjoint path for multi-layer protection in GMPLS networks

In this paper, different recovery methods applied at different network layers and time scales are used in order to enhance the network reliability. Each layer deploys its own fault management methods. However, current recovery methods are applied to only a specific layer. New protection schemes, based on the proposed partial disjoint path algorithm, are defined in order to avoid protection duplications in a multi-layer scenario. The new protection schemes also encompass shared segment backup computation and shared risk link group identification. A complete set of experiments proves the efficiency of the proposed methods in relation with previous ones, in terms of resources used to protect the network, the failure recovery time and the request rejection ratio

Enhanced multi-layer protection in multi-service GMPLS networks

This paper focuses on QoS routing with protection in an MPLS network over an optical layer. In this multi-layer scenario each layer deploys its own fault management methods. A partially protected optical layer is proposed and the rest of the network is protected at the MPLS layer. New protection schemes that avoid protection duplications are proposed. Moreover, this paper also introduces a new traffic classification based on the level of reliability. The failure impact is evaluated in terms of recovery time depending on the traffic class. The proposed schemes also include a novel variation of minimum interference routing and shared segment backup computation. A complete set of experiments proves that the proposed schemes are more efficient as compared to the previous ones, in terms of resources used to protect the network, failure impact and the request rejection ratio