Semantic Web and Social Network

Semantic Web and Social Network
Dr.AstonCole Profile Pic
Dr.AstonCole,United Kingdom,Researcher
Published Date:10-07-2017
Your Website URL(Optional)
Comment
Social Networks and the Semantic Web Peter ´ Mika December 18, 2006SIKS Dissertation Series No. 2007-03. The research reported in this thesis has been carried out under the auspices of SIKS, the Dutch Graduate School for Information and Knowledge Systems. Promotiecommissie: prof.dr. J. M. Akkermans (promotor, VUA/FEW) prof.dr. T. Elfring (promotor, VUA/FSW) dr. P. Groenewegen (co-promotor, VUA/FSW) prof.dr. P. A. A. van den Besselaar (Universiteit van Amsterdam) prof.dr. G. M. Duysters (Technische Universiteit Eindhoven) prof.dr. J. A. Hendler (University of Maryland) prof.dr. J. Kleinnijenhuis (Vrije Universiteit Amsterdam) prof.dr. A. Th. Schreiber (Vrije Universiteit Amsterdam) prof.dr. B. J. Wielinga (Universiteit van Amsterdam) Copyright° c 2006 by Peter MikaVRIJE UNIVERSITEIT Social Networks and the Semantic Web ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad Doctor aan de Vrije Universiteit Amsterdam, op gezag van de rector magnificus prof.dr. L.M. Bouter, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de faculteit der Sociale Wetenschappen op maandag 5 februari 2007 om 10.45 uur in de aula van de universiteit, De Boelelaan 1105 door Peter ´ Mika geboren te Boedapest, Hongarijepromotoren: prof.dr. J.M. Akkermans prof.dr. T. Elfring copromotor: dr. P. GroenewegenPreface ”A tie is anything you can tell a story about.” - Harrison White It was a rainy winter day on February 13, 2004 when I arrived with a group of PhD students to the Dutch ski-resort town of Bergen (lit. mountains). We all signed up for what has been only referred to inside the university as the super-AiO course: two-days of training for PhD candidates (AiOs) in the art and science of graduating successfully and on time. Our trainers, Brigitte and Jeanine have gone fairly confident about their work: after all, they had given the course before dozens of times. There was one difference this time, namely that all of us were foreigners and the course has been given in English for the first time. Nevertheless, our trainers spoke fluent English as most residents of Holland do, and thus expected few problems in dealing with our somewhat more diverse group. Most of the first day has been spent with fun exercises such as directing each other blind-folded through the forest around the conference center. (An exercise in team build- ing and communication, where failing and falling coincide.) This has lifted our spirits considerably and we didn’t mind when something more serious came along: a lesson in planning, supposedly the skill most obviously in lack by most PhD students. In particular, this was a part in the program where our trainers were trying to bring through the message that one can only reach his or her goals in life by drawing up some kind of a master plan. So we were asked to draw an image of where we would like to be in thirty years from now. (Example: famous professor with lots of free time.) We had to reason backwards from there, i.e. where do you need to be in ten years to get there etc., the idea being that if you follow this chain of reasoning, you can arrive to three concrete steps you can take right now to get closer to your goals in thirty years time. They would ask us to write down these steps on a postcard; they would mail these cards to us in three months time so that we would be confronted with the (lack of the) fulfilment of our promises. The reaction from our group was a strong and immediate form of protest that took our trainers utterly by surprise. And then and there was the moment when I also realized what binds us together beyond our differences: the experience that life is not something you plan. In fact, probably the only thing we had in common was that at one point or another we all took a step into the unknown. We have decided to pursue a PhD in the Netherlands,leaving friends and family behind. A carefully considered step in the right direction: yes. With fully predictable consequences: hardly. We didn’t deny the usefulness of some planning. And we are not necessarily adventurers either. But based on our life experi- ence, all of us seemed to agree that a life that follows a master plan without unforeseen risks and sidetracks is not one we would like to live. In June 2004, the word ‘serendipity’ has been voted as one of the top ten English words that are hardest to translate. Nevertheless, it comes closest to describe the experi- ence of finding valuable things not by explicitly looking, but rather by generally walking along the right path, keeping an open mind and yes, having some luck sometimes. Partly chosen, partly given, I feel that serendipity has served me well in the past five years. The road I’ve chosen took me through a one-year master’s program, a half-year internship at Aidministrator (now Aduna) and four years of PhD at the ever unofficial Semantic Web Group of the VUA. Along the way I have met a wonderful group of people who supported me, guided me and contributed to every step I made. I have learned that science of any kind is first and foremost about people, and it may not be incidental that this is also a subject of this thesis... First is my second family of friends and colleagues. I owe them that I can now honestly say: I would not have liked to spend my past years in any other way. I’ll not list your names at the risk of forgetting some... I will miss you. My only consolation is that (as my own website reminds me) our networks in science are well able to extend across geographical and organizational boundaries (and if Skype doesn’t work, you can always come to visit me). Speaking of my thesis, it certainly would not have been made possible without my first promotor, Hans Akkermans. He put a great deal of trust in me, while looking out for me all along. I’m grateful in particular for giving me the freedom to explore the Semantic Web realm before settling on the topic of this thesis. As expected from a supervisor, he shared his significant experience in conducting research, being a researcher and dealing with researchers. However, it is through his personality, above and beyond all, that he conveyed the most important meta-lessons of research: that you should not be afraid to fail and that progress is made by looking beyond the conventional, the ordinary and the boring. Tom Elfring and Peter Groenewegen joined Hans in my supervision once I started on my interdisciplinary research within the newly established VU Research School for Business Information Sciences (VUBIS). They have been my guides through the won- derland of Social Science, a world remarkably different from my home base in Computer Science. It’s been a true exchange with Tom and Peter. I’m most grateful for what I’ve learned, for all their effort in trying to understand the technological details of my work and for working with me to translate the outcomes into concrete and relevant results in the field of applied network analysis. In terms of the number of discussions, a special mention goes to Frank van Harmelen. I’ve counted them and Frank is the person I’ve exchanged the most emails with in the past years. And then I didn’t count the number of discussions on the corridor, during lunch or traveling to or from somewhere. Simply, he is the kind of role model that every PhD student should have in the vicinity.Lastly, there are important people in my personal life I would like to acknowledge here. First, my mother who passed away too early to see me graduate. Then the rest of my family who had to miss me for long periods and then had to be satisfied with seeing me for only a couple of days at a time: kosz ¨ on ¨ om, ¨ hogy mellettem alltatok ´ es ´ seg´ ıtettetek I will never be able to thank enough Andor, who undoubtedly made the most personal sacrifices for my PhD. Last, but not least I owe Dirk the world for his unwavering love in the past years. And hopefully for more to come... Peter ´ Mika Amsterdam December 15, 2006Contents 1 Introduction 1 1.1 Social Networks and the Semantic Web . . . . . . . . . . . . . . . . . 2 1.2 Research Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3 Research Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4 Relevance to Social Science . . . . . . . . . . . . . . . . . . . . . . . 5 1.5 Relevance to Information and Computer Science . . . . . . . . . . . . 6 1.6 Structure of this Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.7 Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 I The Semantic Web and Social Networks 9 2 The Semantic Web 11 2.1 Questions and answers . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.1.1 What’s wrong with the Web? . . . . . . . . . . . . . . . . . . . 12 2.1.2 Diagnosis: A lack of knowledge . . . . . . . . . . . . . . . . . 17 2.2 The semantic solution . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.3 Key concepts of the Semantic Web . . . . . . . . . . . . . . . . . . . . 20 2.4 Development of the Semantic Web . . . . . . . . . . . . . . . . . . . . 23 2.5 The emergence of the social web . . . . . . . . . . . . . . . . . . . . . 28 2.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3 Social Network Analysis 33 3.1 What is network analysis? . . . . . . . . . . . . . . . . . . . . . . . . 33 3.2 Development of Social Network Analysis . . . . . . . . . . . . . . . . 35 3.3 Key concepts in network analysis . . . . . . . . . . . . . . . . . . . . . 37 3.3.1 The global structure of networks . . . . . . . . . . . . . . . . . 38 3.3.2 The macro-structure of social networks . . . . . . . . . . . . . 44 3.3.3 Personal networks . . . . . . . . . . . . . . . . . . . . . . . . 47 3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52ii Contents II Semantic Technology for Social Network Analysis 53 4 Electronic sources for network analysis 55 4.1 Electronic discussion networks . . . . . . . . . . . . . . . . . . . . . . 56 4.2 Blogs and online communities . . . . . . . . . . . . . . . . . . . . . . 57 4.3 Web-based networks . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 5 Ontology-based Knowledge Representation 67 5.1 The Resource Description Framework (RDF) and RDF Schema . . . . . 68 5.1.1 RDF and the notion of semantics . . . . . . . . . . . . . . . . . 71 5.2 The Web Ontology Language (OWL) . . . . . . . . . . . . . . . . . . 73 5.3 Comparison to the Unified Modelling Language (UML) . . . . . . . . . 75 5.4 Comparison to the Entity/Relationship (E/R) model and the relational model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 5.5 Comparison to the Extensible Markup Language (XML) and XML Schema 81 5.6 Discussion: Web-based knowledge representation . . . . . . . . . . . . 85 6 Modelling and aggregating social network data 87 6.1 State-of-the-art in network data representation . . . . . . . . . . . . . . 88 6.2 Ontological representation of social individuals . . . . . . . . . . . . . 90 6.3 Ontological representation of social relationships . . . . . . . . . . . . 94 6.3.1 Conceptual model . . . . . . . . . . . . . . . . . . . . . . . . 96 6.4 Aggregating and reasoning with social network data . . . . . . . . . . . 102 6.4.1 Representing identity . . . . . . . . . . . . . . . . . . . . . . . 103 6.4.2 On the notion of equality . . . . . . . . . . . . . . . . . . . . . 104 6.4.3 Determining equality . . . . . . . . . . . . . . . . . . . . . . . 106 6.4.4 Reasoning with instance equality . . . . . . . . . . . . . . . . . 108 6.4.5 Evaluating smushing . . . . . . . . . . . . . . . . . . . . . . . 111 6.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 6.5.1 Advanced representations . . . . . . . . . . . . . . . . . . . . 112 7 Implementation of the methods 115 7.1 Developing Semantic Web applications with social network features . . 117 7.1.1 Sesame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 7.1.2 Elmo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 7.1.3 GraphUtil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 7.2 Flink . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 7.2.1 The features of Flink . . . . . . . . . . . . . . . . . . . . . . . 124 7.2.2 System design . . . . . . . . . . . . . . . . . . . . . . . . . . 126 7.3 openacademia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 7.3.1 The features of openacademia . . . . . . . . . . . . . . . . . . 131 7.3.2 System design . . . . . . . . . . . . . . . . . . . . . . . . . . 134 7.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139Contents iii III Case Studies 141 8 Evaluating electronic data extraction for network analysis 143 8.1 Differences between survey methods and electronic data extraction . . . 145 8.2 Context of the empirical study . . . . . . . . . . . . . . . . . . . . . . 147 8.3 Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 8.4 Preparing the data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 8.5 Optimizing goodness of fit . . . . . . . . . . . . . . . . . . . . . . . . 150 8.6 Comparison across methods and networks . . . . . . . . . . . . . . . . 153 8.7 Predicting the goodness of fit . . . . . . . . . . . . . . . . . . . . . . . 154 8.8 Evaluation through analysis . . . . . . . . . . . . . . . . . . . . . . . . 158 8.9 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 9 Semantic-based Social Network Analysis in the sciences 163 9.1 Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 9.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 9.2.1 Data acquisition . . . . . . . . . . . . . . . . . . . . . . . . . 166 9.2.2 Representation, storage and reasoning . . . . . . . . . . . . . . 169 9.2.3 Visualization and Analysis . . . . . . . . . . . . . . . . . . . . 170 9.2.4 Electronic data for network analysis . . . . . . . . . . . . . . . 171 9.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 9.3.1 Descriptive analysis . . . . . . . . . . . . . . . . . . . . . . . 174 9.3.2 Structural and cognitive effects on scientific performance . . . . 176 9.4 Conclusions and Future Work . . . . . . . . . . . . . . . . . . . . . . . 182 10 Ontologies are us: emergent semantics in folksonomy systems 183 10.1 A tripartite model of ontologies . . . . . . . . . . . . . . . . . . . . . . 184 10.1.1 Ontology enrichment . . . . . . . . . . . . . . . . . . . . . . . 186 10.2 Case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 10.2.1 Ontology emergence in del.icio.us . . . . . . . . . . . . . . . . 188 10.2.2 Community-based ontology extraction from Web pages . . . . . 193 10.3 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 10.4 Conclusions and Future Work . . . . . . . . . . . . . . . . . . . . . . . 195 IV Conclusions 197 11 The perfect storm 199 11.1 Looking back: the story of Katrina PeopleFinder . . . . . . . . . . . . 200 11.1.1 The Semantic Web . . . . . . . . . . . . . . . . . . . . . . . . 203 11.1.2 Social Networks . . . . . . . . . . . . . . . . . . . . . . . . . 207 11.2 Looking ahead: a Second Life . . . . . . . . . . . . . . . . . . . . . . 209 Samenvatting 213 Summary 217iv Contents Bibliography 221 SIKS Dissertation Series 229Chapter 1 Introduction Modern day research is faced with both extraordinary opportunities and challenges. On the one hand, a fast paced modern society turns to academics as public servants for immediate answers to the practical problems created by its own increasing needs and desires. Society is willing to invest in research as the basis of a knowledge economy as long as research proves to be responsive to its needs. On the other hand, most of the questions science is required to answer are too com- plex to be addressed in the traditional disciplinary framework of academic research. Yet with the explosion of knowledge, research has become only more fragmented than ever. The lack of communication even within the faculties of a single university leads to op- portunities lost every single day. (And even research groups within universities become specialized: long gone are the days where every subdiscipline within a scientific domain was equally represented at a university.) Most apparent is the way the extraordinary technological developments of the com- puter age are left confined to the area of Computer Science. A few successes such as the combination of Biology and Informatics in the field of Bioinformatics show us a glimpse of the unprecedented capabilities of a multidisciplinary approach in addressing problems that have been thought to be too difficult only a decade ago due to the human effort in- volved. Replacing human effort with computing power led to discoveries such as the map of the human genome. And such interaction is not just a one-way technology trans- fer from Computer Science to an application area: information science itself is shaped increasingly by ideas borrowed from application areas such as the biological world. The research contained in this thesis answers some specific research questions in Computer Science and Social Science and proves that an interdisciplinary approach brings appropriate results and contributes to our understanding of both fields. On the one hand we address theoretical questions and practical problems in Social Network Analysis by using the methods of Computer Science in the process of data collection, management and presentation. On the other hand, we apply the world view and methods of Network Analysis in understanding the role of social networks in the technology that underlies the Semantic Web, a technological innovation that will lead us to a next generation of the World Wide Web.2 Introduction This work would not have been possible without a new form of academic research funding at the Vrije Universiteit Amsterdam (VUA), which instigates interdisciplinary research and puts the necessary structures in places. In 2003, the VU has begun imple- menting its new long term research vision by providing special funding for interdiscipli- nary research. The Vrije Universiteit Research School for Business Information Sciences (VUBIS) has come into existence as a collaboration between the Faculties of Science (FEW), Social Science (FSW), Economics and Business Administration (FEWEB). The VUBIS initiative obtained initial funding for eight interdisciplinary PhD projects, includ- ing the one discussed in this thesis. 1.1 Social Networks and the Semantic Web The Semantic Web is a term coined by Tim Berners-Lee for the next stage in the evolution of the World Wide Web he initiated. In this vision of the future Web, information is given well-defined meaning (semantics) in a way that allows our computers to combine and reason with information from multiple sources just as we do ourselves when we search and browse the Web. Since our machines have only limited ways to access the semantics of information at the moment (to understand the content of text, images etc.), additional formal descriptions need to be provided for the information and services populating the web. For the sake of interoperability such descriptions have to be expressed in shared, formal and domain-specific vocabularies: these ontologies capture the agreement within a community over the set of concepts and relationships in the domain and contain logic- based descriptions of these. In order to prevent the segmentation of the Semantic Web into islands of semantics, users and communities are also given the possibility to use and extend each other’s ontologies, forming a Web of ontologies and metadata. Although there is still a scarcity of semantics on the Web, the idea of the Semantic Web has inspired a new stream of research of applying the results of Knowledge Representation in the setting of the Web as well as in other scenarios (e.g. enterprise systems). In our primary study, we apply Semantic Web technology to the aggregation of the electronic data sets that we collect about the social networks of researchers working on the realization of the Semantic Web. We analyze these data using the methods of network analysis and make contributions to the field of scientometrics by measuring the impact of social networks on the success (or failure) of researchers. In our secondary study, we look at the role of social networks within the architecture of the Semantic Web. Although this has been largely overlooked by the early proponents of the Semantic Web, it is now apparent that the Semantic Web is as much a socio- technological innovation as a purely technological one. While the expectation was that the Semantic Web would function with highly formal ontologies with minimal ambiguity and thus a minimal need for human interpretation, we now encounter the limitations of increasing the formality of knowledge in the global, dynamic environment of the Web. As a result, much of the semantics in the lightweight ontologies we find is not part of formal agreements but implicit in the way ontological terms are used by a certain community of users. In particular, the folksonomies or tagging systems that are the basis1.2. Research Questions 3 of many novel knowledge sharing applications developed under the banner of Web 2.0 are too poor to be understood in the logical framework of the Semantic Web. The methods of network analysis on the other hand can be effectively applied to recover the shared, implicit meaning behind the terms in folksonomies by looking at their usage patterns within certain communities. 1.2 Research Questions We propose the following two independent research questions, each with a number of sub-questions. ² How does the relational and cognitive structure of social networks affect the sub- stantial outcomes of emerging scientific-technological communities such as the Semantic Web? – Can we exploit the Web as a data source for social network analysis? – How could we assess the reliability of network data obtained from the Web? – How could we support the reuse and aggregation of electronic data in com- plex studies in Social Network Analysis? ² What is the computational role of social networks in the emergence of semantics in folksonomies? – Can we conceptualize folksonomies as lightweight, dynamic, socially grounded ontologies? – If yes, how can we extract the semantics of terms emerging through usage? The primary research question is relevant to the Social Sciences, in particular to un- derstanding the effects of social networks on innovation and science. Our sub-questions are motivated by the technological opportunities obtaining large scale data for network analysis from the Web and applying Semantic Web technologies in the management of social network data. The secondary research question is relevant to Information and Computer Science, in particular to the development of new forms of emergent ontologies for Semantic Web. It is motivated by the close tie between social networks and cognitive similarity and builds on available data in the form of large scale lightweight social-semantic networks (folksonomies). 1.3 Research Methodology Answering interdisciplinary queries such as the ones posed above requires an extended world view and approach. In terms of world view, the researcher needs to transcend the boundaries of single disciplines and acquire knowledge of the different domains of investigation (information systems on the one hand and social systems on the other) and4 Introduction their connectivity. Based on this new world view existing methods of investigation may also need to be combined into a new, interdisciplinary methodology. In disciplinary works of Social Science, the objects of study are real world objects; in the case of network analysis the focus is on the social ties and social groups that make up the structure of human communities. Two phases of research constitute the accepted methodology of the research field. In the first phase of research development the researcher is required to generate hypotheses about social structures by observing them in the real world. The dominating methods for this kind of study are qualitative, including direct observation (field study) and unstructured interviews. In the second phase of research development, the researcher is required to verify the hypothesis by testing it using quantitative methods, in the same or different context. It is common that the two phases of research are executed by different persons, in particular the same hypothesis is typically tested in different settings before it becomes an accepted part of scientific knowledge. In comparison, the world of Information and Computer Science is populated by in- formation artifacts; the fact that some of these information artifacts may represent real world objects (e.g. an eight-by-eight matrix of bytes used to represent a chess table) is of limited concern to the information sciences: once an informational representation of a real world phenomenon is established (i.e. we agree that an eight-by-eight matrix of bytes captures all important aspects of a chess puzzle), the link between the model and the real world can be ignored. The task of the researcher is typically the engineering of abstract methods (algorithms) that transform information artifacts in desirable ways, e.g. apply the rules of chess to come up with a solution to the puzzle. The properties of the algorithms are also of interest, e.g. whether a solution is found for all puzzles with a solution (completeness) or what computational properties they have (e.g. time and space complexity.) In this thesis we take an interdisciplinary, holistic view of the world when treating both of our research questions. In this view the informational world is not a separate distillate of the real world but constituted by it. This also means that the link between the two cannot be ignored. In fact, the link between the two worlds serves as a major inspiration for our work. As we have observed in our work, acknowledging this world view takes a leap of faith from the disciplinary researcher. This is most visible where domains and methods cross disciplinary domains and these points can be easily identified. In terms of Social Science, the leap of faith concerns the use of online data in place of real world data. While we return to this point later, we note that even today researchers in this area feel compelled to cite the early work of Wellman Wellman et al., 1996, which for them sanctions the use of electronic data such as data from the Web. Discussions about the particulars such as how well the Web reflects publication networks are still de- bated, however. The caution towards online data is least prevalent in the methodological 1 core of network analysis community , but more strongly felt in the application areas of network analysis such as Organization Science and Management Science. As we might 1 Sessions about the analysis of online networks have been part of the International Sunbelt Social Network Analysis conference series at least since 1998. The Sunbelt community itself has also embraced the use of the internet early on with a regular mailing list, educational website and online journal.1.4. Relevance to Social Science 5 expect, this is no concern to the computer scientists (like the author of this thesis), who are glad to mine the Web for real world knowledge. In terms of Computer Science, the leap of faith is required to understand the role of social networks as part of the Semantic Web architecture. In particular, computer scien- tists are predisposed to a view of knowledge as an abstract artifact that can be detached from the social context. While again we return to this argumentation later, we note here that the developments in the last year of this PhD point to a growing acceptance of the 2 role of social networks. 1.4 Relevance to Social Science 1. Theoretical contributions (a) This thesis provides a methodology based on Semantic Web technology for supporting the full process of extracting, representing, aggregating, storing and visualizing social network data. Semantic Web technology plays a partic- ular role in supporting the aggregation of data from heterogeneous sources of information, leading to more robust and more easily comparable results in network analysis. (b) Further, we explore the possibility of applying Web mining to social network data acquisition from the Web in order to support large scale, longitudinal studies in network analysis. We improve on existing methods and for the first time provide an evaluation of web-based extraction in comparison to the survey method of data collection most commonly used in Social Network Analysis. (c) Lastly, we provide theoretical contributions by applying our methodology to the field of scientometrics. Based on a large scale, multi-source data set we test the positive impacts of a structurally and cognitively diverse personal networks as measured by the performance of researchers within the Semantic Web community. 2. Practical results (a) Both the implementation of our methods and the data set collected in this thesis are available as open source for conducting further experiments in the same or different domains. We demonstrate our results through the Flink website, which displays the social networks and research profiles of members of the Semantic Web research community. As an application of Semantic Web technology, Flink has been awarded a first prize at the Semantic Web 3 Challenge of 2004. 2 In particular, the International Semantic Web Conference (ISWC) in 2005 has seen a number of workshops dedicated to the topic as well as a session devoted to social networks and the Semantic Web. At the same time, the EU has also supported the startup of a significant project on this topic with close to 50 million euros of funding. 3 Seehttp://challenge.semanticweb.org6 Introduction 1.5 Relevance to Information and Computer Science 1. Theoretical contributions (a) Driven by the requirements of data aggregation in the science domain, we ex- plore the general problem of data heterogeneity at the instance level. While the more common methods of ontology mapping target the mapping of on- tologies on the schema level, the goal of instance unification or smushing is to map instances that denote the same real world entity. We explore the requirements of smushing from the representation side and discuss ways to use existing tools (query engines and reasoners) for performing the mapping. (b) Based on the observation that social networks influence how we conceptu- alize the world, we further explore the social dimension of semantics in a study of folksonomies, lightweight semantic structures where the seman- tics of terms is largely implicit in their usage. We propose a representation of folksonomies that incorporates the actor who is making annotations and show how methods of network analysis can be used to uncover the seman- tics emerging through usage. We also show that the perceived correctness of emergent models is in fact dependent on the social context. This work has been awarded a Best Paper award at the International Semantic Web Confer- ence (ISWC) of 2005. 2. Practical results (a) The implementation of our methods for instance unification is available as part of the open source Elmo package, an API for the popular Sesame 4 ontology storage facility. 1.6 Structure of this Thesis Every thread in this thesis leads to one or more of the three major studies described in Chapters 8, 9 and 10. In Chapter 8, we compare the results of novel methods of social network extraction from the Web with the outcomes of a survey in a research community. We use data from social network mining and other sources in our scientometric study of the Semantic Web research community in Chapter 9. We explore the role of the social context in knowledge representation for the Semantic Web in Chapter 10. Chapters 4, 5, 6 and 7 contain the necessary details that would allow anyone to re- produce our work and to design and execute social studies where data are collected auto- matically and aggregated using semantic technology. Chapters 2 and 3 introduce the key concepts of the Semantic Web and Social Network Analysis. These chapters prepare both computer scientists with an interest in networks and social scientists with an interest in electronic data for understanding the subsequent discussions. 4 Seehttp://www.openrdf.org1.7. Publications 7 1.7 Publications This thesis is based on and has led to the following list of international refereed publica- tions: ² Hans Akkermans and Peter Mika. Ontology Technology, Knowledge Articulation, and Web Innovation, chapter in: Advances in Knowledge Management Vol. III, Jos Schreinemakers et al. (eds.), Ergon Verlag, 2006 ² Peter Mika. Ontologies are us: A unified model of social networks and semantics. Journal of Web Semantics 4 (4), 2006. To Appear. ² Peter Mika, Tom Elfring and Peter Groenewegen. Application of semantic tech- nology for social network analysis in the sciences. Scientometrics 68 (1), page 3–27, 2006 ² Peter Mika. Ontologies are us: A unified model of social networks and seman- tics. In: Proceedings of the Fourth International Semantic Web Conference (ISWC 2005), Yolanda Gil, Enrico Motta, Richard V. Benjamins and Mark Musen (eds.) , Lecture Notes in Computer Science no. 3729, page 122–136, Galway, Ireland, November, 2005. Winner of the Best Paper Award at ISWC 2005. ² Peter Haase, Bjorn Schnizler, Jeen Broekstra, Marc Ehrig, Frank van Harmelen, Maarten Menken, Peter Mika, Michal Plechawski, Pawel Pyszlak, Ronny Siebes, Steffen Staab and Christoph Tempich. Bibster – A Semantics-Based Bibliographic Peer-to-Peer System. Journal of Web Semantics 2 (1), page 99–103, 2005 ² Peter Mika. Flink: Semantic Web Technology for the Extraction and Analysis of Social Networks. Journal of Web Semantics 3 (2), page 211–223, 2005. Winner of the Semantic Web Challenge of 2004. ² Peter Mika. Social Networks and the Semantic Web: The Next Challenge. IEEE Intelligent Systems 20 (1), January/February, pages 80–93, 2005 ² Peter Haase, Jeen Broekstra, Marc Ehrig, Maarten Menken, Peter Mika, Michal Plechawski, Pawel Pyszlak, Bjorn Schnizler, Ronny Siebes, Steffen Staab and Christoph Tempich. Bibster – A Semantics-Based Bibliographic Peer-to-Peer Sys- tem. In: Proceedings of the Third International Semantic Web Conference (ISWC 2004), Sheila A. McIlraith, Dimitris Plexousakis and Frank van Harmelen (eds.), page 122–136, Hiroshima, Japan, November, 2004 ² Peter Mika. Social Networks and the Semantic Web: An Experiment in Online Social Network Analysis. In: Proceedings of the IEEE/WIC/ACM International Conference on Web Intelligence, pages 285–291, Beijing, China, September, 2004 ² Peter Mika. Bootstrapping the FOAF-web: An Experiment in Social Network Mining. In: 1st Workshop on Friend of a Friend, Social Networking and the Se- mantic Web, Galway, Ireland, September, 20048 Introduction ² Peter Mika and Aldo Gangemi. Descriptions of Social Relations. In: Proceedings of the 1st Workshop on Friend of a Friend, Social Networking and the (Semantic) Web, 2004 ² Peter Mika and Hans Akkermans. Towards a New Synthesis of Ontology Technol- ogy and Knowledge Management. Knowledge Engineering Review 19 (4), page 317–345, 2004 ² Peter Mika, Marta Sabou, Aldo Gangemi and Daniel Oberle. Foundations for DAML-S: Aligning DAML-S to DOLCE. In: Proceedings of First International Semantic Web Services Symposium (SWS2004), AAAI Spring Symposium Series, 2004 ² Peter Mika, Daniel Oberle, Aldo Gangemi and Marta Sabou. Foundations for Service Ontologies: Aligning OWL-S to DOLCE. In: Proceedings of the 13th International World Wide Web Conference (WWW2004), pages 563–572, 2004 ² Aldo Gangemi and Peter Mika. Understanding the Semantic Web through De- scriptions and Situations. In: On The Move 2003 Conferences (OTM2003), Robert Meersman, Zahir Tari and Douglas Schmidt et al. (eds.), pages 689–706, 2003 ² Peter Mika, Victor Iosif, York Sure and Hans Akkermans. Ontology-based Con- tent Management in a Virtual Organization, chapter in: Handbook on Ontologies in Information Systems, Steffen Staab and Rudi Struder (eds.) , International Hand- books on Information Systems, page 447–471, 2003 ² Christiaan Fluit, Herko ter Horst, Jos van der Meer, Marta Sabou and Peter Mika. Spectacle, chapter in: Towards the Semantic Web: Ontology-Driven Knowledge Management, ISBN 0-470-84867-7, 2003 ² Victor Iosif, Peter Mika, Rikard Larsson and Hans Akkermans. Field Experiment- ing With Semantic Web Tools In A Virtual Organization, chapter in: Towards the Semantic Web: Ontology-Driven Knowledge Management, ISBN 0-470-84867-7, 2003 ² Peter Mika. Integrating Ontology Storage and Ontology-based Applications Through Client-side Query and Transformations. In: Proceedings of Evaluation of Ontology-based Tools (EON2002) workshop at EKAW2002, Siguenza, Spain, 2002 ² Peter Mika. JavaServer Pages (In Hungarian.), chapter in: The J2EE Guide for Java Programmers, J. N. Gaizler (eds.), page 115–163, ISBN 963-463-578-4, 2002

Advise: Why You Wasting Money in Costly SEO Tools, Use World's Best Free SEO Tool Ubersuggest.