How the Semantic Web Is Being Used: an Analysis of FOAF
Total Page:16
File Type:pdf, Size:1020Kb
1 How the Semantic Web is Being Used: An Analysis of FOAF Documents Li Ding¤, Lina Zhouy, Tim Finin¤, Anupam Joshi¤ ¤ Department of Computer Science and Electrical Engineering y Department of Information Systems University of Maryland, Baltimore County Baltimore MD USA fdingli1,zhoul,finin,[email protected] Abstract— Semantic Web researchers have initially focused on engineering (more than 1,000 RDF documents has defined the representation, development and use of ontologies but paid terms containing ‘person’ 2). The other well populated non- less attention to the social and structural relationships involved. meta ontologies in Table I include: DC (Dublin Core Element The past year has seen a dramatic increase in the amount of published RDF documents using the Friend of a Friend (FOAF) Set) [3], which defines document metadata properties without vocabulary, providing a valuable resource for investigating how domain/range qualification, and RSS (RDF Site Summary), early Semantic Web adopters use this technology as well as which is “a lightweight multipurpose extensible metadata build social networks. We describe an approach to identify, description and syndication format” for annotating websites discover, and analyze FOAF documents. Over 1.5 million of [4] 3. FOAF documents are collected to show the variety and scalability of the web of FOAF documents. We analyzed the empirical usage TABLE I of namespace and properties in the FOAF community, which helps the FOAF project in standardizing vocabularies. We also EIGHT BEST POPULATED ONTOLOGIES (GENERATED IN JUNE,2004) analyzed the social networks induced by those FOAF documents Onto. Namespace URI # of Docs. and revealed interesting patterns which can become powerful Name Populated resource for outsourcing and justification of scientific knowledge. RDF http://www.w3.org/1999/02/22-rdf-syntax-ns# > 1; 129; 749 FOAF http://www.foaf-project.org/ > 1; 126; 002 DC http://purl.org/dc/elements/1.1/ > 1; 117; 433 I. INTRODUCTION RDFS http://www.w3.org/2000/01/rdf-schema# > 1; 129; 749 MCVB http://webns.net/mvcb/ > 8; 838 The Semantic Web offers a promising solution to publishing RSS http://purl.org/rss/1.0/ > 7; 560 information and services on the World Wide Web augmented vCard http://www.w3.org/2001/vcard-rdf/3.0# > 6; 229 with descriptions in a form that is easier for machines to Bio http://purl.org/vocab/bio/0.1/ > 6; 183 process and understand. This will help Web agents to per- forming a variety of tasks on behalf of their users, such as FOAF provides an RDF/XML vocabulary to describe per- information discovery and integration and service negotiation sonal information [5], including name, mailbox, homepage and composition. Information published in the Semantic Web URL, friends, and so on. FOAF documents then induces the languages (RDF and OWL) uses terms denoting classes and “web of acquaintances” [6] and thus an implicit trust network properties drawn from one or more ontologies. These ontolo- to support such applications as knowledge outsourcing [7] and gies are online RDF documents that declare a set of terms online communities [8]. with unique URIs and further define them by asserting logical The advances in FOAF vocabulary and applications high- relationships and constraints among them. Among a large number of ontologies that have been pub- light several challenging issues that must be addressed. For lished on the Web, however, only a few are well populated, example, how can one assemble a collection of FOAF doc- i.e., have any significant use. 1 Our recent investigation of uments to support Semantic Web research? What are the the namespaces of well populated ontologies (see Table I) common patterns of connections among FOAF documents? revealed that, besides the meta-level ontologies (i.e. RDF, What terms in FOAF vocabulary are the most frequently used? RDFS, DAML and OWL), one of the best populated ontology What is the potential of FOAF in enabling and enhancing the is FOAF (Friend-of-a-Friend) [1]. In addition, representing intelligence of Web-based information systems? The current personal information is also a popular theme in ontology FOAF literature ([9], [5], [8], [10], [11], [6]) provides a vision and various models of how FOAF documents might be used to Partial research support was provided by DARPA contract F30602-00- support Web-based information system under the assumption 0591 and NSF awards ITR-IIS-0326460 and ITR-IIS-0325464. We greatfully that FOAF documents are widely available. There is still a acknowledge many contributions from and interactions with colleagues in the UMBC ebiquity research group. 1When a term in an ontology is used (e.g., an instance of a class is created, 2This is reported by our Swoogle (http://swoogle.umbc.edu), a RDF crawl- or a property used to assert a relationship), we say that the term (and its ing and indexing engine [2]. ontology) is being populated. This is similar to the use of populate to refer 3The use of RSS is increasing dramatically as of this writing and Swoogle to adding actual data to a database. has discovered approximate 80,000 RSS documents by September, 2004. 2 lack of an empirical investigation on the characteristics and The remainder of this paper is organized as follows. Section structure of the growing body of millions of FOAF documents. two presents a review of the literature concerning FOAF vo- This paper presents the first empirical result to answer the cabulary and social network analysis. Section three introduces above questions based on a large collection (over 1.5 million) a novel approach to building FOAF documents collection of real world FOAF documents harvested from the Web. and analyzing the structure of friendship networks in the Our research on online FOAF documents consists of four Semantic Web. Section four uses descriptive statistics and steps: identification of FOAF documents, discovery of FOAF social network analysis to present findings on components of documents using software agents, extraction of person in- FOAF documents and structural relationships among person formation, and fusion of person information based on the profiles. Section five concludes with a discussion the findings semantics of FOAF vocabulary. Using the statistics over this of this study and their implications to the Semantic Web collection of FOAF documents, we describe the common research and practice. properties and namespaces shared by the FOAF community. We hope that this analysis might help FOAF developers design II. BACKGROUND and build better tools as well as inform novice FOAF users A. FOAF Document on how to create effective FOAF documents. Analyses of the A FOAF document publishes “Web homepages for people, social networks encoded in FOAF documents provides insight groups, companies and other kinds of thing”, and it is “written into some interesting structural patterns of the Semantic Web in XML syntax, and adopts the conventions of the Resource from the person perspective. The richness of profiles in FOAF Description Framework (RDF)” [16]. The FOAF project [1] documents allows us to further characterize social ties and was initiated by Dan Brickley and Libby Miller. It enriches identify friendship types. the expression of personal information and relationships. So A direct result of this study will be a friendship directory it is a useful building block for creating information systems in the Semantic Web. Based on this directory, reputation and that support online communities [5]. recommendation systems can be maintained to help people The most important component of a FOAF document is choose trustworthy information (or service) providers and the FOAF vocabulary, which is identified by the namespace propagate trust through friendship relations. For example, URI http://xmlns.com/foaf/0.1/. The FOAF vocabulary defines an agent helping a customer find a good but inexpensive both classes (e.g., foaf:Agent, foaf:Person, and foaf:Document) restaurant in a region might sort the recommendations based and properties (e.g., foaf:name, foaf:knows, foaf:interests, and on the distance to the recommender in the social network. foaf:mbox) grounded in RDF semantics. In contrast to a fixed Friendship networks connected by FOAF relationships can standard, the FOAF vocabulary is managed in an open source provide insights into features and patterns of social networks manner, i.e., it is not stable and is open for extension [1] 4. in the Semantic Web and advance the theories and models of Therefore, inconsistent FOAF vocabulary usage is expected social structures. Friendship networks in the physical world across different FOAF documents. Currently, a large amount have been long studied in the social science. A well known of FOAF documents are contributed by the fast-growing ‘blog’ example is Milgram’s small-world phenomenon [12] – the websites. observation that everyone in the world can be reached through The practical significance of FOAF to information creators a short chain of social acquaintances. The concept gives rise and consumers can be illustrated with a variety of applications to the famous phrase six degrees of separation, which has re- [5], [10], which are summarized as follows: cently been applied to social network analysis in both physical To creators, FOAF is useful by and virtual environments (e.g., [13], [9]). Social relationship have been derived from the contextual information or domain ² Managing communities by offering a basic expression knowledge (e.g. co-citation relationship [14]) indirectly using for community membership. Many communities have data mining techniques. In addition to social networks, the proliferated on the Web, ranging from companies through collection of FOAF documents can serve as valuable resource professional organizations to social groups. for Semantic Web research in the development and testing of ² Expressing identity by allowing unique user IDs across trust models as well as trust propagation models [15]. applications and services without compromising privacy.