Service Matchmaking and Resource Retrieval in the Semantic Web ...

2 downloads 259 Views 3MB Size Report
Oct 27, 2008 - best service or the best resource when searching in a web of meanings like the. Semantic ..... the SAWSDL-MX matching engine returns a ranked list of service offers that ...... to the device hosting a discovered profile as β. 1.
The 7th International Semantic Web Conference

Ruben Lara Tommaso Di Noia Ioan Toma

Service Matchmaking and Resource Retrieval in the Semantic Web (SMR2 2008) October 27, 2008

The 7th International Semantic Web Conference October 26 – 30, 2008 Congress Center, Karlsruhe, Germany

Platinum Sponsors

Ontoprise

Gold Sponsors

BBN eyeworkers Microsoft NeOn SAP Research Vulcan

Silver Sponsors

ACTIVE ADUNA Saltlux SUPER X-Media Yahoo

The 7th International Semantic Web Conference October 26 – 30, 2008 Congress Center, Karlsruhe, Germany

Organizing Committee General Chair

Tim Finin (University of Maryland, Baltimore County) Local Chair

Rudi Studer (Universität Karlsruhe (TH), FZI Forschungszentrum Informatik) Local Organizing Committee

Anne Eberhardt (Universität Karlsruhe) Holger Lewen (Universität Karlsruhe) York Sure (SAP Research Karlsruhe) Program Chairs

Amit Sheth (Wright State University) Steffen Staab (Universität Koblenz Landau) Semantic Web in Use Chairs

Mike Dean (BBN) Massimo Paolucci (DoCoMo Euro-labs) Semantic Web Challenge Chairs

Jim Hendler (RPI, USA) Peter Mika (Yahoo, ES) Workshop chairs

Melliyal Annamalai (Oracle, USA) Daniel Olmedilla (Leibniz Universität Hannover, DE) Tutorial Chairs

Lalana Kagal (MIT) David Martin (SRI) Poster and Demos Chairs

Chris Bizer (Freie Universität Berlin) Anupam Joshi (UMBC) Doctoral Consortium Chairs

Diana Maynard (Sheffield) Sponsor Chairs

John Domingue (The Open University) Benjamin Grosof (Vulcan Inc.) Metadata Chairs

Richard Cyganiak (DERI/Freie Universität Berlin) Knud Möller (DERI) Publicity Chair

Li Ding (RPI) Proceedings Chair

Krishnaprasad Thirunarayan (Wright State University) Fellowship Chair

Joel Sachs (UMBC)

Preface Welcome to the International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web, SMR2 , Karlsruhe, Germany, October 27, 2008. The workshop, at its second edition, is devoted to the discussion of theoretical, technical and methodological solutions to the problem of finding the best service or the best resource when searching in a web of meanings like the Semantic Web is. This year we accepted 8 papers covering many aspects of matchmaking and retrieval in the Semantic Web. As in the last edition, many of them focus on (semantic) web service discovery and selection. In Semantic Web Service Selection with SAWSDL-MX, Matthias Klusch and Patrick Kapahne extend and adapt their hybrid -MX matchmaking framework to SAWDL semantic web service descriptions. SAWSDL is also the main character of the paper Uncovering WSDL Specifications’ Data Semantics by George A. Vouros et al. Here the focus is on automatic annotation of WSDL specifications mapping input/output WSDL specifications to ontology classes. Whenever a service matchmaker returns a list of service satisfying a specific goal the main question is: how satisfactory is the result with respect to the provided goal? An answer to this basic question is provided in Evaluating Semantic Web Service Matchmaking Effectiveness Based on Graded Relevance (Ulrich K¨ uster and Birgitta K¨ onig-Ries) where the authors propose a graded relevance scale to evaluate SWS matchmakers. In all interoperability scenarios, the ultimate goal of matchmakers is the eventual orchestration of discovered services. In Model-Driven Semantic Service Matchmaking for Collaborative Business Processes, Matthias Klusch et al. propose to apply the principles of model driven-design to Semantic Web service technology to assist a business orchestrator finding suitable services at design time, and composing work-flows for agent-based execution. Usually, a negotiation phase follows the matchmaking/discovery one. In Combining Boolean Games with the Power of Ontologies for Automated Multi-Attribute Negotiation in the Semantic Web, Thomas Lukasiewicz and Azzurra Ragone propose a new formal framework to combine Semantic Web technologies with a game theoretic approach for multi-attribute negotiation. Talking about Semantic Web we do not have to forget that related technologies can also be applied in scenarios different from the Web. In Match’n’Date: Semantic Matchmaking for Mobile Dating in P2P Environments, Michele Ruta et al. describe an application of Semantic Web technologies to a mobile environment. Finally, in Look Ma, No Hands: Supporting the semantic discovery of services without ontologies (George A. Vouros et al.) and Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques (Alberto Fernandez et al.) the authors show how to use and combine techniques and tools of the current web to solve problems in the Semantic Web.

Our thanks go to all authors for their valuable submissions and to the invited speaker Holger Lausen for his talk: Enabling Discovery of Web Services on the Internet. We are also very grateful to the members of the Program Committee and the external reviewers for their time and efforts. Tommaso Di Noia, Ruben Lara and Ioan Toma

SMR2 PC chairs

Workshop Organization Program co-chairs Tommaso Di Noia (Technical University of Bari, Italy) Ruben Lara (Telef´ onica I&D, Spain) Ioan Toma (STI Innsbruck, Austria)

Steering Committee Abraham Bernstein (U. Zurich, Switzerland) Tommaso Di Noia (TU Bari, Italy) Takahiro Kawamura (Toshiba, Japan) Matthias Klusch (DFKI, Germany) Ulrich Kster (U. Jena, Germany) Ruben Lara (Telefonica R&D, Spain) Alain Leger (France Telecom, France) David Martin (SRI International, USA) Terry Payne (U. Southampton, UK) Axel Polleres (DERI, National University of Ireland, Galway) Massimo Paolucci (NTT DoCoMo Europe, Germany) Ioan Toma (STI Innsbruck, Austria)

Program Committee Sudhir Agarwal (University of Karlsruhe, Germany) Rama Akkiraju (IBM, USA) Sinuh Arroyo (U. Alcala de Henares, Spain) Djamal Benslimane (Universit Claude Bernard Lyon, France) Gheorghe Cosmin Silaghi (Babes-Bolyai University Cluj-Napoca, Romania) Eugenio Di Sciascio (Technical University of Bari, Italy) Stephan Grimm (FZI Karlsruhe, Germany) Sung-Kook Han (Won Kwang University, Korea) Frank Kaufer (University of Potsdam, Germany) Uwe Keller (STI Innsbruck, Austria) Holger Lausen (seekda, Austria) Freddy Lecue (Orange-France Telecom, France) Ioan Alfred Letia (Technical University of Cluj-Napoca, Romania) Christophe Rey (ISIMA, University of Clermont Ferrand, France) Dumitru Roman (STI Innsbruck, Austria) Farouk Toumani (ISIMA, University of Clermont Ferrand, France)

External Reviewers Claudio Baldassarre (The Open University, UK) Claudia d’Amato (University of Bari, Italy) Alessio Gugliotta (The Open University, UK) Jacek Kopecky (University of Innsbruck, Austria) Tomasz Kaczmarek (Poznan University of Economics, Poland)

Table of Contents Invited Talk: Enabling Discovery of Web Services on the Internet .................................................... 1 Holger Lausen

Semantic Web Service Selection with SAWSDL-MX ........................................................................... 3 Matthias Klusch and Patrick Kapahne Uncovering WSDL Specifications' Data Semantics ............................................................................ 19 George A. Vouros, Alexandros Valarakos, Konstantinos Kotis Evaluating Semantic Web Service Matchmaking Effectiveness Based on Graded Relevance ........ 35 Ulrich Küster and Birgitta König-Ries Model-Driven Semantic Service Matchmaking for Collaborative Business Processes .................... 51 Matthias Klusch, Stefan Nesbigall, and Ingo Zinnikus Combining Boolean Games with the Power of Ontologies for Automated Multi-Attribute Negotiation in the Semantic Web .......................................................................................................... 67 Thomas Lukasiewicz and Azzurra Ragone Match'n'Date: Semantic Matchmaking for Mobile Dating in P2P Environments ........................... 83 Michele Ruta, Tommaso Di Noia, Eugenio Di Sciascio, and Floriano Scioscia Look Ma, No Hands: Supporting the semantic discovery of services without ontologies ................ 99 George A. Vouros, Fragkiskos Dimitrokallis, Konstantinos Kotis Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques .......... 115 Alberto Fernandez, Conor Hayes, Nikos Loutas,Vassilios Peristeras, Axel Polleres, Konstantinos Tarabanis

1

Invited Talk: Enabling Discovery of Web Services on the Internet Holger Lausen seekda OG Museumstrae 21/302a – 6020 Innsbruck Austria

The Web is moving from a collection of static documents to a set of Web Services. Todays major search engines provide fast and easy access to existing Web pages, however only little attention has been paid to provide a similar easy and scalable access to find existing publicly available Web Services. We present an approach that considers existing practical realities and has been used to build the seekda.com Web Service search engine. Using this approach seekda has indexed the largest pool of Web Service known so far. The talk will give details on how existing Web Service related data can be obtained from the Web, how it can be analyzed to obtain semantic annotations, how availability monitoring can be used to assure accuracy and finally ideas on how user feedback can be used to improve the quality of the available information.

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

2

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

3

Semantic Web Service Selection with SAWSDL-MX Matthias Klusch and Patrick Kapahnke German Research Center for Artificial Intelligence Stuhlsatzenhausweg 3, Saarbr¨ ucken, Germany [email protected], [email protected]

Abstract. In this paper, we present an approach to hybrid semantic Web service selection of semantic services in SAWSDL based on logicbased matching as well as text retrieval strategies. We discuss the principles of semantic Web service description in SAWSDL and selected problems for service matching implied by its specification. Based on the result of this discussion, we present different variants of hybrid semantic selection of SAWSDL services implemented by our matchmaker called SAWSDL-MX together with preliminary results of its performance in terms of recall/precision and average query response time. For experimental evaluation we created a first version of a SAWSDL service retrieval test collection called SAWSDL-TC.

1

Introduction

As a W3C recommendation dated August 28, 2007, the SAWSDL1 specification proposes mechanisms to enrich Web services described in WSDL2 (Web Service Description Language) with semantic annotations. However, there is no SAWSDL semantic service matchmaker publicly available to the community yet. To fill this gap, we initially adopt the ideas of semantic Web service matching of our hybrid matchmakers OWLS-MX and WSMO-MX (see [8, 6]), for service description languages OWL-S3 and WSML respectively, to this environment. A detailed discussion of the SAWSDL specification, particularly addressing the problems arising for semantic Web service selection, is also given. In this paper, we present the first version of our hybrid SAWSDL Web service matchmaker called SAWSDL-MX. It exploits both crisp logic-based matching (subsumption reasoning) and IR-based (text retrieval) matching. Our preliminary experimental analysis shows, that in line with OWLS-MX and WSMO-MX, hybrid matching can outperform both variants applied stand-alone in terms of recall and precision. The remainder of this paper is structured as follows. After a brief introduction to SAWSDL and discussion of implied challenges of semantic service selection in 1 2 3

http://www.w3.org/TR/sawsdl/ http://www.w3.org/TR/wsdl/ and http://www.w3.org/TR/wsdl20/ http://www.daml.org/services/owl-s/1.1/

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Semantic Web Service Selection with SAWSDL-MX

4

section 2, the hybrid matching approach of SAWSDL-MX is described in detail in section 3. Section 4 presents the architecture and implementation details of SAWSDL-MX. Preliminary results of our experimental evaluation of SAWSDLMX over a initial test collection SAWSDL-TC1 in terms of recall, precision and average query response time are shown in 5. We comment on related work in section 6 and conclude in section 7.

2

SAWSDL Services

In the following, a brief introduction of the semantically enabled service description language SAWSDL is given. Language specific problems for semantic service discovery arising from the W3C recommendation and methods of resolution and assumptions for avoiding them respectively are also discussed. SAWSDL is designed as extension of WSDL enabling service providers to enrich their service descriptions with additional semantic information. For this purpose, the notion of model reference and schema mapping have been introduced in terms of XML attributes that can be added to already existing WSDL elements as depicted in figure 1. More precisely, the following extensions are used for annotation:

Fig. 1. SAWSDL extensions of WSDL interface components

– modelReference: A modelReference points to one ore more concepts with equally intended meaning expressed in an arbitrary semantic representation language. They are allowed to be defined for every WSDL and XML Schema element, though the SAWSDL specification defines their occurrence only

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Semantic Web Service Selection with SAWSDL-MX

in WSDL interfaces, operations, faults as well as XML Schema elements, complex types, simple types and attributes. The purpose of a model reference is mainly to support automated service discovery. – liftingSchemaMapping: Schema mappings are intended to support automated service execution by providing rules specifying the correspondences between semantic annotation concepts defined in a given ontology (the ”upper” level) to the XML Schema representation of data actually required to invoke the Web service using SOAP (the ”lower” level), and vice versa. A liftingSchemaMapping describes the transformation from the ”lower” level in XML Schema up to the ontology language used for semantic annotation. – loweringSchemaMapping: The reference tag loweringSchemaMapping describes the transformation from the ”upper” level of a given ontology to the ”lower” level in XML Schema. Since the specification of SAWSDL does not restrict the developer of a semantic service in SAWSDL to a particular ontology language, any service selection has to cope with the implied semantic interoperability problem of both heterogeneous ontologies and heterogeneous ontology languages. Therefore, as an initial starting point, we restricted our inital SAWSDL service matchmaker to ”understand” only the standard OWL4 . More concrete, we assume for SAWSDL-MX 1.0 that model references in SAWSDL service offers and requests are pointing to ontological concepts exlcusively defined in OWL-DL. That allows to apply standard subsumption reasoning used for OWL-S matchmaking such as in [14, 4, 8]. Besides, there is no retrieval test collection for SAWSDL publicly available yet, but for OWL-S, namely OWLS-TC, which we converted semi-automatically into SAWSDL services such that we could use the resulting SAWSDL-TC for initially evaluating our matchmaker. Another problem with the SAWSDL specification with respect to service matching is that so-called top-level annotation and bottom-level annotation are defined as to be considered independent from each other. The term top-level annotation describes the case, where a complex type or element definition of a message parameter is described by a model reference as a whole. A bottom-level annotation pursues the idea of semantically annotating the parts that are contained inside the definition of a complex type or element. However, it remains unclear how to evaluate matching between top-level and low-level annotated parameters, or which one to prefer if both levels are available. To circumvent this problem, we decided to rely on top-level annotations of upper parameter type definitions, and ignore bottom-level annotation in the first version of our matchmaker. In addition to that, element and type definition specifying a message component can be annotated at the same time. The specification does not imply a solution for this case either, so we decided to rely on the annotation directly attatched to the referenced XML Schema object if available. Further, multiple references to multiple ontologies defined in different languages and formats such as logic theories, plain text documents or structured 4

http://www.w3.org/2004/OWL/

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

5

6

Semantic Web Service Selection with SAWSDL-MX

Fig. 2. SAWSDL service example

thesauri can be used to describe the semantic of even the same element. Therefore, a matchmaker, in principle, cannot know whether these different types of semantic descriptions of the element are intended to be treated as complementary or equivalent. In the first case, how to aggregate the complementing descriptions, in the latter case, which one to select best for further processing? This opens up a wide range of pragmatic approaches to deal with this for service matching. SAWSDL-MX 1.0 checks only the first model reference of an element. However, different variants dealing with multiple model references connected to a single object are topic of further development, since they are to be treated as sets without order. One possible approach would be to check every combination of request and service offer reference part and perform some kind of aggregation afterwards. To illustrate this problem by example, consider figure 2: A flight company offers a WSDL Web service with different operations concerning flight booking (BookFlight operation), account administration (omitted in the picture), and so on. The BookFlight operation is defined to take information of the desired flight (Flight input) and customer information in form of a tuple containing a user name and appropriate password (Customer input) as input parameters and delivers information about the ticket reservation (Ticket output). To support automated Web service selection, this service is semantically annotated in compliance with the SAWSDL specification as shown in the figure.

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Semantic Web Service Selection with SAWSDL-MX

In particular, the service developer of the flight company uses WSML-Core, WSML-Rule and OWL-DL concept descriptions for service element annotation. As a consequence, a matchmaker agent cannot perform single language-specific reasoning and matching mechanisms but has to apply an appropriate combination of them instead. This problem can be straight-forwardly solved by use of language mappings available for WSML-Core and OWL-DL5 but remains hard to solve for comparing concepts in WSML-Rule and OWL-DL. Further, in the example, the XML Schema description attached to the service input element Customer contains annotations for the compound complex type as well as the simple types (referenced by the elements contained in the complex type, element nodes are omitted in the picture). How to handle this situation? Selecting only one annotation level may neglect additional information while looking at all references as a conjunction of ontological concepts can lead to either logical inconsistencies, or is not possible due to incomparable description languages. This problem is exaggerated in the example by providing multiple references (multiple levels of annotations) for the same element. SAWSDL-MX 1.0 only checks the top-level annotation of the most generic element of the XML Schema description of a service parameter.

3

Service Matching with SAWSDL-MX

In the following, we describe one approach to SAWSDL-service selection which we implemented in an initial version of a matchmaker called SAWSDL-MX based on the assumptions stated above. SAWSDL-MX performs service selection in terms of logic-based, syntactic (text similarity-based) and hybrid matching of I/O parameters defined for potentially multiple operations of a Web service interface (signature matching)6 . As service requests, standard SAWSDL Web service definition documents are used. This approach is particularly inspired by the hybrid semantic service matchmakers OWLS-MX [8] and WSMO-MX [6] for OWL-S and WSML. 3.1

Service Interface Matching

The matching process of SAWSDL-MX on the service interface level is performed as follows. For every pair of service offer O and service request R, every combination of their operations is evaluated by either logic-based matching, text 5 6

http://www.wsmo.org/TR/d16/d16.1/v0.21/ For SAWSDL-MX 1.0, we assume only one interface but multiple operations per service. Extending the proposed service matching algorithm to services with even multiple interfaces only requires additionally combined valuation of the respective interface matching results. The restriction to signature matching for SAWSDL-MX 1.0 is due to the fact that, in SAWSDL, preconditions and effects can be added as input and output model references only, which makes it hard for any matchmaker to identify them as such in general, and before actually analyzing the name and content of referenced models in particular.

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

7

8

Semantic Web Service Selection with SAWSDL-MX

retrieval-based matching, or both. The matching of operations is described in more detail later. In order to compute an optimal injective mapping of operations for service offer and request, SAWSDL-MX applies bipartite graph matching, where nodes in the graph represent the operations and the weighted edges are built from possible one-to-one assignments with their weights derived from the computed degree of operation match. If there exists such a mapping, then it is guaranteed that there exists an operation provided by the service offer for every operation a requester defined in her query. That is, there exists no request operation that cannot be provided by the service offer, disregarding the quality of match at this point. As an example, consider the service request and service offer given in figure 3. Every request operation ROi (with i ∈ {1, 2}) is compared to every advertisement operation Oj (with j ∈ {1, 2, 3}) with respect to logic-based filters defined in the next section. In this example, RO1 exactly matches with O1 , but fails for O2 and O3 . O3 is a weaker plug-in match for RO2 (the subsumed-by match of RO2 with O2 is even weaker than a plug-in match). The best (max) assignment of matching operations is {RO1 , O1 , RO2 , O3 }.

Fig. 3. Interface level matching of SAWSDL-MX

One conservative (min-max) option of determining the matching degree between service offer and request based on their pairwise operation matchings is to assume the worst result of the best operation matchings, to guarantee a fixed lower bound of similarity for every requested operation. This is what

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Semantic Web Service Selection with SAWSDL-MX

SAWSDL-MX 1.0 is doing, so in this example shown in figure 3, the service offer is considered a plug-in match for the request. Other possibilities are to merge the operation matching results based on, for example, their average syntactic similarity values, and to provide more detailed feedback to the user on the operation matchings involved. Please note that SAWSDL-MX aims at finding service matches solely based on single service offer documents. The problem of semantic Web service composition is somehow related, but additional state-based planning strategies have to be applied to solve this problem, which is out of the scope of this work. To accomplish on that, a Web service composition planner like e.g. OWLS-XPlan or SHOP2 could be considered (see [18, 19] for details). 3.2

Logic-based Operation Matching

As mentioned above, we assume for SAWSDL-MX 1.0 that model references in SAWSDL service offers and requests are pointing to ontological concepts exlcusively defined in OWL-DL or WSML-DL. That allows to apply standard subsumption reasoning for description logics (see [20]). Therefore, the logic-based operation matching part of SAWSDL-MX computes the degree of logic-based match for a given pair of service offer operation OO and service request OR by successively applying four filters of increasing degree of relaxation: Exact, Plugin, Subsumes and Subsumed-by, which are, in essence, adopted from those of OWLS-MX 2.0 but modified in terms of an additional bipartite concept matching to ensure an injective mapping between offer and request concepts, if required. The reason of this modification is that previous experiments with OWLS-MX showed that many logic-based only failures could have been avoided by this additional constraint. Exact match: Service operation OO exactly matches service operation OR ⇔ (∃ injective assignment Min : ∀m ∈ Min : m1 ∈ in(OO ) ∧ m2 ∈ in(OR ) ∧ m1 ≡ m2 ) ∧ (∃ injective assignment Mout : ∀m ∈ Mout : m1 ∈ out(OR ) ∧ m2 ∈ out(OO ) ∧ m1 ≡ m2 ). There exist a one-to-one mapping of perfectly matching inputs as well as perfectly matching outputs. Assuming that an operation fullfills a requesters need if every input can be satisfied and every requested output is provided, the assignments only require to be injective (but not bijective), thus additional available information not required for service invocation and additional provided outputs not explicitly requested are tolerated. Plug-in match: Service operation OO plugs into service operation OR ⇔ (∃ injective assignment Min : ∀m ∈ Min : m1 ∈ in(OO )∧m2 ∈ in(OR )∧m1 m2 )∧ (∃ injective assignment Mout : ∀m ∈ Mout : m1 ∈ out(OR ) ∧ m2 ∈ out(OO ) ∧ m2 ∈ lsc(m1 )). The filter relaxes the constraints of the exact matching filter by additionally allowing input concepts of the service offer to be arbitrarily more general than those of the service request, and advertisement output concepts to be direct child concepts of the queried ones. Subsumes match: Service operation OO subsumes service operation OR ⇔ (∃ injective assignment Min : ∀m ∈ Min : m1 ∈ in(OO ) ∧ m2 ∈ in(OR ) ∧ m1

m2 ) ∧ (∃ injective assignment Mout : ∀m ∈ Mout : m1 ∈ out(OR ) ∧ m2 ∈

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

9

Semantic Web Service Selection with SAWSDL-MX

10

out(OO ) ∧ m1 m2 ). This filter further relaxes constraints by allowing service offer outputs to be arbitrarily more specific than the request outputs (as opposed to the plug-in filter, where they have to be direct children). Thus, a plug-in can be seen as special case of a subsumes match resulting in a more fine-grained view at the overall service ranking. Subsumed-by match: Service operation OO is subsumed by service operation OR ⇔ (∃ injective assignment Min : ∀m ∈ Min : m1 ∈ in(OO ) ∧ m2 ∈ in(OR ) ∧ m1 m2 ) ∧ (∃ injective assignment Mout : ∀m ∈ Mout : m1 ∈ out(OR )∧m2 ∈ out(OO )∧m2 ∈ lgc(m1 )). The idea of the subsumed-by matching filter is to determine the service offers that the requester is able to provide with all required inputs and at the same time deliver outputs that are at least closely related to the requested outputs in terms of the inferred concept classification. At this filtering step, services that offer equivalent or more specific outputs already have been discovered. The subsumed-by filter additionally returns service offers that provide more general output concepts, namely direct parents. These may be of value for a user to know, though it depends on the granularity of the matchmaker ontology. For example, it would not make sense to return a service operation providing information on vehicles, if the user explicitly requested information on a very special brand of a car which concept is inappropriately modelled as a direct child of the concept vehicles in the ontology. The overall algorithm for logic-based matching of operations considers the filters in the following order based on the degree of relaxation: exact > plug-in > subsumes > subsumed-by > fail. The notion of fail applies to cases where none of the filtering tests succeeded. 3.3

Syntactic Operation Matching

In addition, SAWSDL-MX can perform syntactic-based matching based on selected token-based text similarity measures. That is, a syntactic similarity value is computed for every pair of service offer and request operation which is used to rank operations with same logic-based matching degree. The implemented similarity measures for SAWSDL-MX 1.0 are the same as for OWLS-MX, that are the Loss-of-Information, the Extended Jaccard, the Cosine and the Jensen-Shannon similarity measures. The architecture of SAWSDL-MX allows the integration of other text similarity measures such as those provided by SimPack7 which is also used in the iMatcher matchmaker [7]. The weighted keyword vectors of inputs and outputs for every operation are generated by first unfolding the referenced concepts in the ontologies (as defined for standard tableaux reasoning algorithms). The resulting set of primitive concepts of all input concepts of a service operation is then processed to a weighted keyword vector based on TFIDF weighting scheme, the same is done with its output concepts. The text similarity of a service offer operation and a request operation is the average of the similarity values of their input and output vectors according to the selected text similarity measure. 7

http://www.ifi.uzh.ch/ddis/research/semweb/simpack/

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Semantic Web Service Selection with SAWSDL-MX

3.4

Hybrid Operation Matching

Inspired by OWLS-MX [8], SAWSDL-MX combines logic-based and syntacticbased matching to perform hybrid semantic service matching. There are different options of combination: A compensative variant using syntactic similarity measures in cases where none of the logic-based filters applies helps to improve the service ranking with respect to logic-based false negatives by re-considering them again in the light of their computed syntactic similarity. An integrative variant deals with problems concerning logic-based false positives by not taking the syntactic similarity of concepts into account only when a logical matching fails, but as a conjunctive constraint in each logical matching filter. Our experiments showed that OWLS-MX 2.0 using the integrative variant performs better than the original one with the complementary use of syntactic similarity. However, SAWSDL-MX 1.0 inherited the compensative variant from OWLS-MX 1.0, that is, only the logic-based subsumed-by filter is modified to a hybrid filter by integrative checking of syntactic simliarity of concepts, and the syntactic nearest-neighbour filter is compensative in the sense that it is only performed in case all other filters fail. Subsumed-by match: Service operation OO is subsumed by service operation OR ⇔ (∃ injective assignment Min : ∀m ∈ Min : m1 ∈ in(OO ) ∧ m2 ∈ in(OR ) ∧ m1 m2 ) ∧ (∃ injective assignment Mout : ∀m ∈ Mout : m1 ∈ out(OR ) ∧ m2 ∈ out(OO ) ∧ m2 ∈ lgc(m1 )) ∧ simIR (OR , OO ) ≥ α. A subsumed-by match computed by hybrid matching additionally requires the IR-based similarity computed using one of the measures from IR = {LOI, ExtJacc, Cos, JS} to be above a given threshold α. This helps to avoid logic-based false positives to be introduced by the pure logic-based variant of this filter. Nearest-neighbour match: This filter compensates logic-based false negatives as described above. Its condition is simIR (OR , OO ) ≥ α and thus considers all services not already catched in previous filter steps whose IR-based similarity is above the threshold.

4

SAWSDL-MX Implementation

SAWSDL-MX 1.0 has been fully implemented in Java using the sawsdl4j8 API (handling SAWSDL for WSDL 1.1) and the OWL API9 for access to SAWSDL and OWL files, the DIG 1.110 as standard interface to handle SHOIQ knowledge base queries, and the Pellet11 reasoner as inference engine for logic-based matchmaking. Figure 4 gives an broad overview of the overall system architecture. Basically, SAWSDL-MX consists of the following components: SAWSDL Matching Engine, Service Registry, Ontology Handlers, Local Matchmaker Ontology and Similarity 8 9 10 11

http://knoesis.wright.edu/opensource/sawsdl4j/ http://owlapi.sourceforge.net/ http://dig.sourceforge.net/ http://pellet.owldl.com/

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

11

12

Semantic Web Service Selection with SAWSDL-MX

Measures. These are described in more detail in the following. From the perspective of service providers, SAWSDL-MX allows the registration of SAWSDL Web service offers at the service registry. For requesters, SAWSDL-MX provides an interface for submitting queries by means of a SAWSDL document specifying details about the desired service interface. After the service discovery process, the SAWSDL-MX matching engine returns a ranked list of service offers that match the query.

Fig. 4. SAWSDL-MX architecture

SAWSDL Matching Engine: The SAWSDL Matching Engine is the core component of SAWSDL-MX. It provides several matching variants of SAWSDLMX 1.0 as described in previous sections: The Logic-based Matcher computes service ranking by means of crisp-logic subsumption reasoning and the logicbased matching filters described in section 3.2. The IR-based Matcher produces the ranked results using syntactic similarity measures as described in section 3.3. Finally, the Hybrid Matcher performs the combined approach of logic-based reasoning and syntactic similarity comparison as described in 3.4. The matching engine component is designed to provide easy integration of additional matching variants by means of Java interface implementation. Service Registry: This component is the storage for service offers provided by service providers. It is accessed by the matching engine to produce the ranked results for a query.

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Semantic Web Service Selection with SAWSDL-MX

Ontology Handlers: After the service registration process, the semantic annotations of a SAWSDL service (by means of model references) are processed using Ontology Handlers. Therefore, an appropriate handler able to parse and reason about the referenced ontology is selected and the concepts are stored locally to facilitate logic-based reasoning as well as concept unfolding for IR-based matching at query time. As for the matching engine component, the Ontology Handlers package is designed to allow the proper integration of additional knowledge representation formalisms by means of Java interfaces. Local Matchmaker Ontology: This component is in fact part of the ontology handlers in the actual implementation but depicted as seperate component for reasons of clarity. The Local Matchmaker Ontology is a storage for all relevant concepts referenced by registered service offers as proposed in [8]. However, since SAWSDL allows the use of various knowledge representation formalisms, parts of the component relevant for certain ontology handlers are directly covered inside the handlers. In case of our current implementation of SAWSDL-MX, it consists of the Pellet reasoner, which is accessed by handlers able to process description logic based ontology languages via DIG 1.1. Currently, only the OWL-DL Handler is actually implemented, but expanding the system to WSML-DL is straight-forward, since they rely on subsets of the SROIQ description language, which is addressed by Pellet12 . Similarity Measures: This package currently contains the four similarity measures loss-of-information, extended Jaccard, cosine and Jensen-Shannon. However, adding more variants for IR-based matching can be easily accomplished again via interfaces. An proprietary document indexing structure based on hash tables is also provided. The integration of additional syntactic similarity measures (e.g. from SimPack) and better indexing strategies is intended for following versions of SAWSDL-MX.

5

Evaluation of Performance

The experimental evaluation of the retrieval performace of the first version SAWSDL-MX focuses on measuring its recall and precision based on a first SAWSDL test collection semi-automatically derived from OWLS-TC 2.213 using the OWLS2WSDL14 tool, as there is currently no standard test collection for SAWSDL matchmaking available. OWLS2WSDL transforms OWL-S service descriptions (and concept definitions relevant for parameter description) to WSDL through syntactic transformation. The collection consists of 894 Web services covering different application domains: education, medical care, food, travel, communication, economy and weaponry. For this set of service offers, 26 queries have been selected and relevance sets have been created for each of them. These where subjectively defined as relevant according to the standard TREC definition of binary relevance [16]. As the creation of this test collection has been done 12 13 14

With exception of n-ary datatypes http://projects.semwebcentral.org/projects/owls-tc/ http://projects.semwebcentral.org/projects/owls2wsdl/

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

13

Semantic Web Service Selection with SAWSDL-MX

14

by transforming OWL-S services contained in OWLS-TC 2.2, which provides services containing only one atomic process per description, every SAWSDL advertisement only contains a single interface with a single operation (but possibly multiple I/O’s). Therefore and because all automatically derived model references exclusively point to OWL ontologies, this test collection can only be seen as a first attempt towards a commonly agreed testing environment for SAWSDL service discovery and our evaluation has to be considered as preliminary. The performance measures used for evaluation are defined as follows: Recall =

|A ∩ B| |A ∩ B| , P recision = , |A| |B|

where A is the set of all relevant documents for a request and B the set of all retrieved documents for a request. The so-called F1-measure equally weights recall and precision and is defined as: F1 =

(2 · P recision · Recall) . (Recall + P recision)

We adopt the prominent macro-averaging of precision. That is, we compute the mean of precision values for answer sets returned by the matchmaker for all queries in the test collection at standard recall levels Recalli (0 ≤ i < λ). Ceiling interpolation is used to estimate precision values that are not observed in the answer sets for some queries at these levels; that is, if for some query there is no precision value at some recall level (due to the ranking of services in the returned answer set by the matchmaker) the maximum precision of the following recall levels is assumed for this value. The number of recall levels from 0 to 1 (in equidistant steps nλ , n = 1 . . . λ) we used for our experiments is λ = 20. Thus, the macro-averaged precision is defined as follows: P recisioni =

 1 max{Po |Ro ≥ Recalli ∧ (Ro , Po ) ∈ Oq }, × |Q| q∈Q

where Oq denotes the set of observed pairs of recall/precision values for query q when scanning the ranked services in the answer set for q stepwise for true positives in the relevance sets of the test collection. For evaluation, the answer sets are the sets of all services registered at the matchmaker which are ranked with respect to their (totally ordered) matching degree. The performance tests have been conducted on a machine with Windows 2000, Java 6, 1,7 GHz CPU and 2 GB RAM using SME2 15 as evaluation environment. As can be seen in figure 5(a), the hybrid variant utilizing cosine measure performs best in both finding correct results among the top of the ranking as well as returning positives at high precision towards full recall. It is followed by pure IR-based service discovery (also using cosine measure), which is surprisingly at first glance, since it is assumed by the semantic Web community that 15

http://projects.semwebcentral.org/projects/sme2/

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Semantic Web Service Selection with SAWSDL-MX

(a) recall/precision

(b) F1

Fig. 5. Performance of SAWSDL-MX

semantically enabled ressource retrieval should be able to outperform standard information retrieval purely relying on syntactic information in general. However, as Wang et al. show in [17] exemplarily for OWL, the currently established Web ontology landscape provides mainly poor specification of concepts in terms of the used expressivity of description languages. In fact, many ontologies currently available are just simple taxonomies that do not rely on advanced features provided by for example OWL-DL, thus IR-based matching techniques are often good enough to compare service parameters. The crisp logic-based variant of SAWSDL-MX performs worst with respect to precision. This is mainly due to the problem with ontologies just described and due to the coarse-grained concept descriptions available. Equal consideration of recall and precision using the F1 measure yields the results given in figure 5(b), which recapitulates the observations. Regarding query response times, IR-based matching performs best, namely 1,7 seconds on average per query, while crisp logic-based matching takes 4,7 seconds on average and their combination in hybrid matching is the slowest (6,4 seconds). These evaluation results are in line with the performance of OWLS-MX and WSMO-MX and thus fortify the proposition that hybrid matching outperforms pure logic-based as well as IR-based matching in terms of recall and precision.

6

Related Work

To the best of our knowledge, there exist only very few implemented semantic service discovery systems for SAWSDL. [10] presents a solution to SAWSDL Web service discovery using UDDI registries called FUSION. In FUSION, any service

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

15

Semantic Web Service Selection with SAWSDL-MX

16

description is classified at the time of its publishing and then mapped to UDDI to allow for fast lookups. In case of unknown semantic service requests reasoning has to be done at query time. In contrast to SAWSDL-MX, each service offer has only to satisfy one matching condition based on subsumption relationships inferred by a reasoner, thus the ranking is not affected by different degrees of logic-based match, neither does FUSION perform a syntactic or hybrid semantic match. Like SAWSDL-MX 1.0, FUSION is strictly bound to OWL-DL, since for each service, a semantic representation in terms of an individual of a predefined OWL concept is constructed. Lumina [11] developed in the METEOR-S project16 follows a similar approach based on a mapping of WSDL-S (and later on SAWSDL respectively) to UDDI but performs syntactic service matching only. For a survey of semantic service matchmakers in general, we refer the interested reader to [9].

7

Conclusion

SAWSDL-MX performs hybrid semantic Web service matching for SAWSDL operations based on both logic-based reasoning and IR-based syntactic similarity measurement, and combines the results to provide a matching result for service interfaces with multiple operations. The requester formulates queries in terms of SAWSDL service interface descriptions and is presented a service ranking containing service offers from the local registry. The version SAWSDL-MX 1.0 presented in this paper has been implemented and evaluated in terms of recall and precision using a preliminary SAWSDL test collection called SAWSDL-TC1 which we derived from the existing collection OWLS-TC 2.2. As the experimental results show, hybrid matching of SAWSDL services can outperform both logicbased and IR-based matching in terms of precision at the cost of increased average query response time. We are currently working on several aspects of SAWSDL service discovery and extensions of SAWSDL-MX. As SAWSDL is not restricted to semantically represent service components using a fixed knowledge representation formalism, the integration of additional ontology language support is intended. While description logics have already been discussed for the first version SAWSDLMX 1.0, the support for languages originating from logic programming such as WSML-Flight and WSML-Rule is subject to our future work. Besides, inspired by the monolithic logic-based semantic service matchmaker MaMaS [1, 2], we are currently working on an adaptive variant called SAWSDLMXA which exploits means of ontology patching such as concept contraction and abduction combined with machine learning based on implicit feedback [5]. The semantic interoperability problem induced by the inevitable occurrence of heterogeneous ontologies used for semantic service annotation can be addressed by appropriate ontology alignment techniques [13]. In SAWSDL-MX, one option is to perform an additional matching of concept primitives (that 16

http://lsdis.cs.uga.edu/projects/meteor-s/

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Semantic Web Service Selection with SAWSDL-MX

are left undefined in the matchmnaker ontology) in unfolded concepts to be compared using a shared minimum vocabulary of requesters and providers like WordNet17 , or by consistent introduction of additional equivalence axioms to the local knowledge base of SAWSDL-MX [12]. SAWSDL-MX 1.0 and SAWSDL-TC1 are both publicly available at semwebcentral.org.

References 1. Colucci, S., Di Noia, T., Di Sciascio, E., Donini, F. M., Mongiello, M.: Concept Abduction and Contraction in Description Logics. Proceedings of the 16th International Workshop on Description Logics (DL’03), Volume 81 - Sept. 2003 2. Colucci, S., Coppi, S., Di Noia, T., Di Sciascio, E., Donini, F. M., Pinto, A., Ragone, A.: SemanticBased Resource Retrieval using Non-Standard Inference Services in Description Logics. Proceedings of Thirteenth Italian Symposium on ADVANCED DATABASE SYSTEMS Sistemi Evoluti per Basi di Dati (SEBD-2005), pp. 232-239, 2005 3. Euzenat, J., Valtchev, P.: Similarity-based ontology alignment in OWL-Lite. Proceedings of the European Conference on Artificial Intelligence ECAI, 333-337, 2004 4. Jaeger, M. C., Rojec-Goldmann, G., Liebetruth, C., M¨ uhl G., Geihs K.: Ranked Matching for Service Descriptions Using OWL-S. KiVS 2005: 91-102, 2005 5. Joachims, T., Radlinski, F.: Search Engines that Learn from Implicit Feedback. Computer Volume 40, Issue 8, Aug. 2007 Page(s):34 - 40, 2007 6. Kaufer, F., Klusch, M.: WSMO-MX: A Logic Programming Based Hybrid Service Matchmaker. Proceedings of the 4th IEEE European Conference on Web Services (ECOWS 2006), IEEE CS Press, Zurich, Switzerland, 2006 7. Kiefer, C., Bernstein, A.: The Creation and Evaluation of iSPARQL Strategies for Matchmaking. Proceedings of the 5th European Semantic Web Conference (ESWC), Lecture Notes in Computer Science, Vol. 5021, pages 463–477, Springer-Verlag Berlin Heidelberg, 2008 8. Klusch, M., Fries, B., Sycara, K.: Automated Semantic Web Service Discovery with OWLS-MX. Proceedings of 5th International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), Hakodate, Japan, ACM Press, 2006 9. Klusch, M.: Semantic Web Service Coordination. In: M. Schumacher, H. Helin, H. Schuldt (Eds.) CASCOM - Intelligent Service Coordination in the Semantic Web. Chapter 4. Birkhuser Verlag, Springer, 2008 10. Kourtesis, D., Paraskakis I.: Combining SAWSDL, OWL-DL and UDDI for Semantically Enhanced Web Service Discovery. Proceedings of the 5th European Semantic Web Conference (ESWC 2008), Lecture Notes in Computer Science (LNCS), vol. 5021, Springer-Verlag Berlin Heidelberg, pp. 614628, 2008 11. Li, K., Verma, K., Mulye, R., Rabbani, R., Miller, J. A., Sheth, A. P.: Designing Semantic Web Processes: The WSDL-S Approach. Chapter submitted to Semantic Web Processes and Their Applications. J. Cardoso, A. Sheth, Editors. Springer 12. Meilecke, C., Stuckenschmidt, H.: Applying Logical Constraints to Ontology Matching. Proceedings of KI 2007: Advances in Artificial Intelligence: 30th Annual German Conference on AI, 2007 13. Shvaiko, P., Euzenat, J.: A Survey of Schema-based Matching Approaches Journal on Data Semantics, 2005. 14. Sycara, K., Paolucci, M., Ankolekar, A., Srinivasan, N.: Automated discovery, interaction and composition of Semantic Web services. Journal of Web Semantics, vol 1, Elsevier, 2003 15. Toch, E., Gal, A., Reinhartz-Berger, I., Dori D.: A Semantic Approach to Approximate Service Retrieval. ACM Transactions on Internet Technology, 8(1), 2008 16. TREC. Text Retrieval Conference. http://trec.nist.gov/data/. 17. Wang, T. D., Parsia, B., Hendler, J.: A survey of the web ontology landscape. Proceedings of International Semantic Web Conference (ISWC), 2006 18. Klusch, M., Gerber, A., Schmidt, M.: Semantic Web Service Composition Planning with OWLSXplan. 1st Intl. AAAI Fall Symposium on Agents and the Semantic Web, Arlington VA, USA, 2005 19. Sirin, E., Parsia, B., Wu, D., Hendler, J., Nau, D.: HTN planning for web service composition using SHOP2. Proceedings of the 2nd International Semantic Web Conference (ISWC), pages 20-23, Sanibel Island, Florida, USA, 2003 20. Baader, F., Calvanese, D., McGuinness, D.L., Nardi, D., Patel-Schneider P.F.: The Description Logic Handbook: Theory, Implementation, and Applications. Cambridge University Press 2003, ISBN 0-521-78176-0 17

http://wordnet.princeton.edu/

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

17

18

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

19

8QFRYHULQJ:6'/6SHFLILFDWLRQV¶'DWD6HPDQWLFV *HRUJH$9RXURV$OH[DQGURV9DODUDNRV.RQVWDQWLQRV.RWLV  $,/DE8QLYHUVLW\RIWKH$HJHDQ.DUORYDVVL6DPRV*UHHFH ^JHRUJHYDOH[YNRWLV`#DHJHDQJU 

$EVWUDFW 7KH ZRUN UHSRUWHG LQ WKLV DUWLFOH DLPV WRZDUGV FRPSXWLQJ ZHE VHUYLFHV¶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

,QWURGXFWLRQ 7KHDXWRPDWLFGLVFRYHU\RI:HEVHUYLFHVLVDKLJKO\LPSRUWDQWVHUYLFHPDQLSXODWLRQ WDVN :6'/ LV PDLQO\ IRFXVLQJ RQ RSHUDWLRQDO DQG V\QWDFWLF GHWDLOV UHJDUGLQJ WKH LPSOHPHQWDWLRQ DQG H[HFXWLRQ RI :HE VHUYLFHV 7KH ODFN RI H[SOLFLW VHPDQWLFV LQ :6'/VSHFLILFDWLRQVPDNHVWKHPLQVXIILFLHQWWRVDWLVI\WKHUHTXLUHPHQWVIRUIOH[LEOH DQG HIIHFWLYH :HE VHUYLFH µPDQLSXODWLRQ¶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¶ VHPDQWLF GHVFULSWLRQ $OWKRXJK RWKHU LPSRUWDQW DVSHFWV H[LVW ZLWK UHVSHFW WR ZHE VHUYLFHV¶ VLJQDWXUH HJ JRDOV SUHSRVW FRQGLWLRQV >@ VHUYLFHV¶ FODVVLILFDWLRQ >@  PDLQO\ GXH WR WKHLU VWURQJ GHSHQGHQFH RQ GDWD VHPDQWLFV WKH\ KDYH QRW UHFHLYHG WKH DWWHQWLRQ WKDW GDWD VHPDQWLFVGR   

 7KLV ZRUN KDV EHHQ VXSSRUWHG E\ WKH (XURSHDQ &RPPLVVLRQ SURMHFW *ULG$OO XQGHU JUDQW QXPEHU)3SURMHFW,67*ULG$OO

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

20

Uncovering WSDL Specifications' Data Semantics

7KHPDLQREMHFWLYHRIWKLVSDSHULVWRVWXG\WKHDXWRPDWLFVHPDQWLFDQQRWDWLRQRI :6'/VSHFLILFDWLRQVJLYHQRQWRORJLHVUHODWHGWRWKHVHUYLFHV¶GRPDLQV7RZDUGVWKLV REMHFWLYH ZH GHYLVH D PHWKRG IRU PDSSLQJ LQSXWRXWSXW PHVVDJHV¶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³H[WHUQDO´ LQIRUPDWLRQ VRXUFHV VXFK DV XVHUV¶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

3UREOHP6WDWHPHQW 7KHSUREOHPWKDWWKLVSDSHUGHDOVZLWKLVDVIROORZV³*LYHQ D D:6'/VSHFLILFDWLRQ DQG E DQRQWRORJ\2  6$ 6EHLQJWKHVHWRIWHUPVOH[LFDOL]LQJRQWRORJ\HOHPHQWV DQG $ WKH RQWRORJ\ D[LRPV SURYLGH PDSSLQJV RI :6'/ PHVVDJHV¶ SDUW QDPHV WR RQWRORJ\FODVVHVZLWKUHVSHFWWRWKHLQWHQGHGVHPDQWLFVRIWKHVHUYLFH´ :H GHDO RQO\ ZLWK VHUYLFHV¶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¶SDUWQDPHVDQGRQWRORJ\HOHPHQWV DUHRIWKHIRUPVHUYLFHBLGPHVVDJHBLGQDPHHOHPHQWUHO!ZKHUHQDPHLVWKHQDPHRI WKHSDUDPHWHUHOHPHQWLVDWHUPLQWKHRQWRORJ\ VLJQDWXUH 6DQG UHOLV WKHDVVHVVHG

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

UHODWLRQEHWZHHQHOHPHQWDQGQDPH7KLVPD\EHHTXLYDOHQFHVXEVXPHVRUVXEVXPHG )RUWKHSXUSRVHVRIWKLVSDSHUZHUHVWULFWRXUDWWHQWLRQWRWKHHTXLYDOHQFHFDVH :H KDYH WR QRWLFH WKDW WKH DERYH SUREOHP VWDWHPHQW IROORZV WKH :6'/  VSHFLILFDWLRQ DOVR DGGUHVVLQJ WKH JUHDW QXPEHU RI H[LVWHQW VHUYLFHV  +RZHYHU WKLV FDQEHHDVLO\UHVWDWHGIRU:6'/VSHFLILFDWLRQVFRQVLGHULQJWKDWQRPHVVDJHSDUW QDPHVH[LVW



5HODWHG:RUNDQG0RWLYDWLRQ

5HODWHG:RUN 7KH ZRUN UHSRUWHG LQ WKLV DUWLFOH DLPV WR SURYLGH VHUYLFHV ZLWK GDWD VHPDQWLFV E\ H[SORLWLQJ RI GRPDLQ RQWRORJLHV VHUYLFHV¶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³VKDOORZ´ IHDWXUHV FRQFHUQLQJ ;0/ VFKHPD DQG RQWRORJ\ HOHPHQWV¶ QDPHV 7KLV IRU LQVWDQFH DIIHFWVPHWKRGV¶HIILFDF\LQFDVHVZKHUHSRO\VHPRXVWHUPVDSSHDURULQFDVHVZKHUH VSHFLILFDWLRQVDUHDWGLIIHUHQWJUDQXODULW\OHYHOVDVIDUDVWKHFRQFHSWXDOL]DWLRQRIWKH GRPDLQLVFRQFHUQHG 7KH ZRUN UHSRUWHG LQ >@ DLPV DW DVVHVVLQJ :6'/ VSHFLILFDWLRQV¶ VLPLODULW\ E\ H[SORLWLQJ WKH VWUXFWXUH RI GDWD W\SHV DQG RSHUDWLRQV RI VHUYLFHV DV ZHOO DV WKH VHPDQWLFV RI QDWXUDO ODQJXDJH GHVFULSWLRQV DQG LGHQWLILHUV $OWKRXJK WKH DLP RI WKLV ZRUNLVWRVXSSRUWTXHU\E\H[DPSOHGLVFRYHU\RIVHUYLFHVWKHHPSKDVLVRQVHPDQWLF PDWFKLQJ JLYHQ WH[WXDO GHVFULSWLRQV RI VHUYLFHV DQG WKH H[SORLWDWLRQ RI LGHQWLILHUV¶ VHPDQWLFV EULQJV WKLV ZRUN FORVH WR RXU ZRUN 7KLV DSSURDFK XVHV D YHFWRU PRGHO H[SORLWLQJWKHWH[WXDOGHVFULSWLRQVRIVHUYLFHVDQGFRQVXOWV:RUG1HWWRFDOFXODWHWKH VHPDQWLFGLVWDQFHVEHWZHHQLGHQWLILHUVRI:6'/HOHPHQWV:KLOHVHPDQWLFPDWFKLQJ LVUHVWULFWHGWRWKHH[SORLWDWLRQRILGHQWLILHUVLGHQWLILHUV¶VHQVHVDUHQRWGLVDPELJXDWHG PDNLQJWKHFDOFXODWLRQRIVHPDQWLFGLVWDQFHVUDWKHUSUREOHPDWLF'LVDPELJXDWLRQLVD YLWDOWDVNLQRXUUHVHDUFKVLQFHZHDLPWRPDWFK:6'/HOHPHQWVWRVSHFLILFRQWRORJ\ FRQFHSWVH[SOLFDWLQJWKHLUVHPDQWLFV $LPLQJ WR VKRZ KRZ FRQWHQWEDVHG DSSURDFKHV FDQ FRQWULEXWH WR VHPDQWLF PDWFKLQJ RI 2:/6 VHUYLFH VSHFLILFDWLRQV 2:/60; >@ DLPV WR H[SORLW WKH LPSOLFLW VHPDQWLFV RI DQ\ SDUW RI 2:/6 VHUYLFH GHVFULSWLRQ E\ UHSUHVHQWLQJ LW DV D

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

21

22

Uncovering WSDL Specifications' Data Semantics

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¶LQSXWRXWSXWSDUDPHWHUVEHIRUHWKHVHDUHEHLQJ FRPSDUHGE\ORJLFEDVHGPDWFKPDNHUV )XUWKHUHYLGHQFHLVSURYLGHGE\H[SHULPHQWVZLWKWKH$66$0¶VDQQRWDWLRQZL]DUG >@ $66$0 FDVWV WKH SUREOHP RI FODVVLI\LQJ RSHUDWLRQV DQG GDWDW\SHV LQ D :HE 6HUYLFH DV D WH[W FODVVLILFDWLRQ SUREOHP 7KH WRRO OHDUQV IURP :HE 6HUYLFHV ZLWK H[LVWLQJVHPDQWLFDQQRWDWLRQV*LYHQWKLVWUDLQLQJGDWDDPDFKLQHOHDUQLQJDOJRULWKP FDQJHQHUDOL]HDQGSUHGLFWVHPDQWLFODEHOVIRUSUHYLRXVO\XQVHHQ:HE6HUYLFHV7KH DSSURDFKGHVFULEHGLQWKHFXUUHQWSDSHUGRHVQRWUHTXLUHWKHSUHH[LVWHQFHRIVHPDQWLF DQQRWDWLRQVWRGHFLGHWKHPDSSLQJRI:6'/HOHPHQWVWRRQWRORJ\FRQFHSWV

0RWLYDWLRQDQGRYHUYLHZRIWKHPHWKRG 2XU ZRUN LV EHLQJ PRWLYDWHG E\ WKH YLHZ WKDW IRU ZHEVHUYLFH PDWFKPDNHUV WR SHUIRUP DFFXUDWHO\ WKH\ QHHG WKH SUHFLVH LQWHQGHG PHDQLQJ RI ZHE VHUYLFHV¶ VLJQDWXUHLQDILQHJUDLQHGZD\7KLVPHDQVWKDWSDUWVRIZHEVHUYLFHPHVVDJHVPXVW EHPDSSHGWRVSHFLILFWHUPVWKDWDUHEHLQJD[LRPDWL]HGLQDIRUPDORQWRORJ\ 7RGRVRZHQHHGWRPLWLJDWHSLWIDOOVUHODWHGWRSKHQRPHQDFRQFHUQLQJV\QRQ\P WHUPV KRPRQ\P WHUPV LH SRO\VHP\  W\SRJUDSKLF YDULDWLRQV RI WHUPV GLIIHUHQFHV EHWZHHQWKHJUDQXODULW\RIVHUYLFHVDQGRQWRORJLFDOVSHFLILFDWLRQV ,QFRQMXQFWLRQWRWKHVHSLWIDOOV ZH QHHGWRWDNHDOVRLQWRDFFRXQWWKH³QDWXUH´RI :6'/VSHFLILFDWLRQVZKLFKDUHEHLQJSURGXFHGIURPSURJUDPFRGHZLWKIHZDQGLQ PRVW RI WKH FDVHV PLVOHDGLQJ FRPPHQWV GHVFULSWLRQV DQG ³WULFN\´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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

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³VHPDQWLF GLVWDQFH´ +RZHYHULQWKLVSDSHUZHRQO\FRQVLGHUFDVHVRIH[DFWPDWFKHV 7R HYDOXDWH 86'6 ZH SURYLGH H[WHQVLYH H[SHULPHQWDO UHVXOWV ZLWK GLIIHUHQW FRQILJXUDWLRQVRIVWULQJPDWFKLQJEDVHGDQGYHFWRUPRGHOEDVHGPHWKRGVLQGLIIHUHQW VHWVRI:6'/VSHFLILFDWLRQV

6HPDQWLF$QQRWDWLRQRI:6'/VSHFLILFDWLRQV $QQRWDWLQJ:6'/ $V SRLQWHG RXW ZH FRQVLGHU WKDW WKH RYHUDOO VHPDQWLF DQQRWDWLRQ RI :6'/ VSHFLILFDWLRQV FRPSULVHV WKUHH GLVWLQFW VWDJHV 7KH DQQRWDWLRQ VWDJH WKH PDSSLQJV VWDJH DQG WKH YDOLGDWLRQ VWDJH +RZHYHU IRU WKH SXUSRVHV RI WKLV SDSHU ZH UHTXLUH KXPDQLQWHUYHQWLRQRQO\LQWKHDQQRWDWLRQVWDJH 

 )LJ$QQRWDWLQJ:6'/VSHFLILFDWLRQV  'XULQJ WKH DQQRWDWLRQ VWDJH KXPDQV SURYLGH WH[WXDO GHVFULSWLRQV IRU HOHPHQWV RI WKH:6'/VSHFLILFDWLRQ7KHDQQRWDWLRQVWDJHWDNHVDVLQSXWWKH:6'/VSHFLILFDWLRQ DQG SURGXFHV DQ H[WHUQDO [PO DQQRWDWLRQ ILOH ($)  EDVHG RQ D VSHFLILF DQQRWDWLRQ

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

23

24

Uncovering WSDL Specifications' Data Semantics

VFKHPD WKDW ZH KDYH VSHFLILHG IRU WKLV SXUSRVH 7KLV LV DQ DQQRWDWLRQWHPSODWH ILOH WKDWSURYLGHV³VORWV´IRUWKHGHVFULSWLRQRI:6'/HOHPHQWV)RUWKHVHUYLFHLWVHOIIRU HDFKLQWHUIDFHRSHUDWLRQDQGLQSXWRXWSXWPHVVDJHV¶HOHPHQWVLWSURYLGHVHOHPHQWVIRU ³FRPPHQWV´ ³GHVFULSWLRQ´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¶ SDUW QDPHV DQG RQWRORJ\ FODVVHV 86'6 LQ DGGLWLRQ WR WKH SDUW QDPHV RI WKH LQSXWRXWSXW PHVVDJHV H[SORLWV VHUYLFHV¶ W\SHV¶ VSHFLILFDWLRQVDVZHOODVWH[WXDOGHVFULSWLRQVDQGFRPPHQWVLQWKHDQQRWDWLRQILOH $VDOUHDG\SRLQWHGRXWWKHFRUHRI86'6FRPSULVHVWKHFRPELQDWLRQRIWZRVWDWH RI WKH DUW PHWKRGV &2&/8 DQG /6$EDVHGPDSSLQJ &2&/8 LV D FRPSUHVVLRQ EDVHGFOXVWHULQJDOJRULWKP &2PSUHVVLRQEDVHG&/8VWHULQJ 7KHDOJRULWKPLVEDVHG RQWKHDVVXPSWLRQWKDWGLIIHUHQWOH[LFDOL]DWLRQVRIDWHUP W\SRJUDSKLFYDULDQWV XVHD FRPPRQVHWRICFRUH FKDUDFWHUV7KHUHIRUHW\SRJUDSKLFYDULDQWVWKDWCPRVWO\ XVHWKLV VHWDUHSRWHQWLDODOWHUQDWLYHOH[LFDOL]DWLRQVRIWKHVDPHFRQFHSWZKLOHWKRVHRQHVWKDW DUHCIDU IURPWKLVVHWDUHSRWHQWLDOO\UHODWHGWRWKHOH[LFDOL]DWLRQRIDGLIIHUHQWFRQFHSW )XUWKHUGHWDLOVIRUWKLVPRGHODQGSDUWLWLRQEDVHGFOXVWHULQJDOJRULWKPDUHSURYLGHGLQ >@&2&/8KDVEHHQXVHG IRUWKHFRPSDULVRQRI D WKHGHVFULSWLRQRIHDFK:6'/ LQSXWRXWSXWPHVVDJHSDUWZLWKWKHODEHOVRIRQWRORJ\FODVVHV E WKHHOHPHQWVLQWKH DWRPLFFRPSOH[ W\SHV VSHFLILFDWLRQV RI PHVVDJHV¶ SDUW HOHPHQWV ZLWK WKH QDPH DQG ODEHOVRIRQWRORJ\FODVVHVDQG F WKHFRPPHQWRIHDFK:6'/LQSXWRXWSXWPDVVDJH SDUWZLWKWKHFRPPHQWVRIRQWRORJ\FODVVHV 7KH /6$EDVHGPDSSLQJ >@ DLPV WR GLVDPELJXDWH WKH PHDQLQJ RI :6'/ PHVVDJHV¶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îP  RI WHUPVGRFXPHQWV 'RFXPHQWV LQ RXU FDVH FRUUHVSRQG WR :RUG1HW VHQVHV 7HUPV DUH VHOHFWHG IURP WKH YLFLQLW\ RI :RUG1HW VHQVHV LHIURPWKHVHQVHVWKHPVHOYHVDQGIURPWKHLUK\SRQ\PK\SHUQ\PV  /DWHQW 6HPDQWLF $QDO\VLV /6$  FRPSXWHV WKH DUUDQJHPHQW RI D NGLPHQVLRQDO VHPDQWLF

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

VSDFHWRUHIOHFWWKHPDMRUDVVRFLDWLYHSDWWHUQVLQWKHGDWD7KLVLVGRQHE\GHULYLQJD VHWRINXQFRUUHODWHGLQGH[LQJIDFWRUVZKLFKPD\EHFRQVLGHUHGDV³ODWHQWFRQFHSWV´ 7KHQHDFKWHUPDQGGRFXPHQWLVUHSUHVHQWHGE\LWVYHFWRURIIDFWRUYDOXHVLQGLFDWLQJ LWVVWUHQJWKRIDVVRFLDWLRQZLWKHDFKRIWKHVH³ODWHQWFRQFHSWV´%\YLUWXHRIGLPHQVLRQ UHGXFWLRQIURPWKH1WHUPVVSDFHWRWKHNIDFWRUVVSDFHZKHUHN1WHUPVWKDWGLGQRW DFWXDOO\ DSSHDU LQ D GRFXPHQW PD\ VWLOO HQG XS FORVH WR WKH GRFXPHQW LI WKLV LV FRQVLVWHQWZLWKWKHPDMRUSDWWHUQVRIDVVRFLDWLRQLQWKHGDWD :KHQRQHVHDUFKHVDQ/6,LQGH[HGGDWDEDVHRIGRFXPHQWVLWSURYLGHVDTXHU\ LH DSVHXGRGRFXPHQW ZKLFKLVDOLVWRIWHUPV7KHVLPLODULW\EHWZHHQWZRGRFXPHQWV LV FRPSXWHG E\ PHDQV RI WKH GRW SURGXFW EHWZHHQ WKH FRUUHVSRQGLQJ UHSUHVHQWDWLRQ YHFWRUV'RLQJVR/6,UHWXUQVDVHWRIJUDGHGGRFXPHQWVDFFRUGLQJWRWKHLUVLPLODULW\ WRWKHTXHU\ ,QRXUFDVHWKHVHPDQWLFVSDFHLVFRQVWUXFWHGE\WHUPVLQWKHYLFLQLW\RIWKHVHQVHV6 6«6P RI WKH :RUG1HW HQWU\ PDWFKLQJ DQ RQWRORJ\ FODVV QDPH & RU D :6'/ PHVVDJHSDUWQDPH&7KHVHWRIWHUPVLQWKHVHPDQWLFVSDFHLQFOXGH  7KHWHUP&¶WKDWFRUUHVSRQGVWR&&¶LVDOH[LFDOHQWU\LQ:RUG1HWWKDWLVD OLQJXLVWLFYDULDWLRQRI&  7HUPVWKDWDSSHDULQ&¶:RUG1HWVHQVHV66«6P  7HUPVWKDWFRQVWLWXWHK\SHURQ\PVK\SRQ\PVRIHDFK&¶VHQVH  7HUPVWKDWDSSHDULQK\SHU K\S RQ\PVRI&¶VHQVHV $VIDUDVRQWRORJ\FODVVHVLVEHLQJFRQFHUQHGHDFKTXHU\LVEHLQJFRQVWUXFWHGE\ H[WUDFWLQJWHUPVIURPWKHODEHODQGWKHFRPPHQWRIDQRQWRORJ\FODVVDVZHOODVIURP WKH QDPHV RI LWV SURSHUWLHV¶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¶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³GLIILFXOW\´RIRXUFDVHVJLYHQWKDW VXFK D VLPSOH PHWKRG IDLOV WR H[KLELW HIIHFWLYH SHUIRUPDQFH7KH /HYHQVKWHLQ PDWFKLQJPHWKRGLQFRUSRUDWHVDGLVWDQFHPHDVXUHWKDWVSHFLILHVWKHPLQLPXPQXPEHU RI RSHUDWLRQV QHHGHG WR WUDQVIRUP RQH VWULQJ LQWR DQRWKHU 7KHYDOLG RSHUDWLRQV DUH LQVHUWLRQGHOHWLRQRUVXEVWLWXWLRQRIDVLQJOHFKDUDFWHU2XULPSOHPHQWDWLRQFRQVLGHUV D PDWFK EHWZHHQ WZR VWULQJV LI DQG RQO\ LI DW PRVW WZR RSHUDWLRQV DUH UHTXLUHG IRU WKHLUWUDQVIRUPDWLRQ

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

25

Uncovering WSDL Specifications' Data Semantics

26

,QDGGLWLRQWRWKHDERYHZHKDYHDOVRXVHGD9HFWRU6SDFH0RGHO±EDVHG 960  PHWKRG ZKLFK FRPSXWHV WKH PDWFKLQJ RI GRFXPHQWV SDLUV (DFK GRFXPHQW LV UHSUHVHQWHG E\ D YHFWRU RI Q ZHLJKWHG LQGH[ WHUPV ,QGH[ WHUPV FRUUHVSRQG WR WKH VLPSOH WHUPV WKDW DUH H[WUDFWHG IURP DOO GRFXPHQWV +HUH ZH FRQVWUXFW SVHXGR  GRFXPHQWVWKDWFRUUHVSRQGWRRQWRORJ\FODVVHVDQG:6'/PHVVDJHV¶SDUWHOHPHQWV $VLWLVGRQHLQWKH/6$EDVHGPHWKRGWKHVHSVHXGRGRFXPHQWVLQFOXGHWHUPVIURP WKH ODEHO DQG WKH FRPPHQW RI DQ RQWRORJ\ FODVV DV ZHOO DV IURP WKH QDPHV RI LWV SURSHUWLHV¶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

'RPDLQ2QWRORJ\ RZO  2QWRORJ\ 7UDYHO 3RUWDO %RRNV 3RUWDO %RRNV &RQFHSW 6802 +RVSLWDO3K\VLFLDQ

FRQFHSWV        

SURSHUWLHV        

 6HUYLFHV :6'          

7RWDO  RI SDUW HOHPHQWV                         

'LVWLQFW SDUW HOHPHQWV WREHDQQRWDWHG         

 7DEOH  VXPPDUL]HV LQIRUPDWLRQ FRQFHUQLQJ WKH LQLWLDO VHWV RI H[SHULPHQWV DQG FKDUDFWHULVWLFV RI WKH VHUYLFHV DQG RQWRORJLHV XVHG 7KH ILUVW DQG WKH VHFRQG FROXPQ SUHVHQW WKH GRPDLQ DQG WKH RQWRORJ\ XVHG 7KH WKLUG DQG IRXUWK FROXPQV SURYLGH LQIRUPDWLRQFRQFHUQLQJWKHQXPEHURIFRQFHSWVDQGSURSHUWLHVLQHDFKRQWRORJ\7KH ILIWKFROXPQSURYLGHVWKHQXPEHURIVHUYLFHV :6'/VSHFLILFDWLRQV WKDWH[LVWLQHDFK

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

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³IUHH WH[W LQFOXGLQJ WHUPV IURP [VGVFKHPD W\SHV DQG IURP WKH :6'/ PHVVDJH SDUW QDPH HOHPHQW´ E $QQRWDWLRQVWKDWDUHIRUPHGE\³IUHHWH[WLQFOXGLQJWHUPVRQO\IURPWKH :6'/ PHVVDJH SDUW QDPH HOHPHQW´ F  $QQRWDWLRQV WKDW DUH IRUPHG E\ ³D VLQJOH WHUP´ $QQRWDWRUV FKRRVH D VLQJOH WHUP ZLWKRXW FRQVLGHULQJ [VGVFKHPD W\SHV RU :6'/ PHVVDJH SDUW QDPH HOHPHQW LQIRUPDWLRQ )RU LQVWDQFH IRU WKH ³ZVGOSDUW QDPH &DSLWDOB&LW\´ WKH DQQRWDWRU FUHDWHV WKH GHVFULSWLRQ DQQRWDWLRQ ³GHVFULSWLRQ! &DSLWDO GHVFULSWLRQ!´ ZKLFK LV WKH LQWHQGHG PHDQLQJ RU V\QRQ\P RIWKHHQWLW\³&DSLWDO&LW\´

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

27

Uncovering WSDL Specifications' Data Semantics

28

7KHUHVXOWRIWKLVSURFHVVLVDVHWRIDQQRWDWLRQVZLWKWHUPVEHORQJLQJLQRQHRIWKH IROORZLQJ WKUHH FDWHJRULHV  D  WHUPV IURP [VGVFKHPD W\SHV E  WHUPV IURP :6'/ PHVVDJHSDUWQDPHVRUF WHUPVFKRVHQE\KXPDQDQQRWDWRUV 7KHXVHRI³IUHHWH[W´LQFRPELQDWLRQZLWK[VGVFKHPDW\SHV¶WHUPVDQGRU:6'/ PHVVDJH SDUW QDPHV¶ WHUPV PHDQV WKDW WKH DQQRWDWRU LV DOORZHG WR IRUP D QDWXUDO ODQJXDJH VHQWHQFH (J ³GHVFULSWLRQ! 7KH VHUYLFH UHTXHVWV DFFRPPRGDWLRQ XVLQJ FRXQWU\LQIRUPDWLRQGHVFULSWLRQ!´$OWKRXJKWKH XVHRIIUHHWH[W PD\GLVWUDFWWKH PDWFKLQJPHWKRGVWKHIUHHGRPWKDWWKHDSSURDFKJLYHVWRWKHDQQRWDWRULVLPSRUWDQW DQG UHDOLVWLF ,Q WKLV H[DPSOH RI DQQRWDWLRQ WKH DQQRWDWRU KDV FRPELQHG WHUPV HJ ³FRXQWU\´ IURPPHVVDJHV¶SDUWQDPHV FDVHE $VDQRWKHUH[DPSOHLQWKHFRPPHQW ³FRPPHQW!7KHFRXQWU\QDPHLVDVWULQJ$FRXQWU\LVGHVFULEHGZLWKLWVFDSLWDOLWV FXUUHQF\DQGLWVJRYHUQPHQWFRPPHQW!´WKHDQQRWDWRUKDVLQFOXGHGWHUPV WHUPV ³FDSLWDO´ ³FXUUHQF\´ ³JRYHUQPHQW´  IURP WKH [VGVFKHPD W\SH WKDW FRUUHVSRQGV WR WKH VSHFLILF PHVVDJH SDUW QDPH WKDW LV DQQRWDWHG FDVH D  7DEOH  VXPPDUL]HV LQIRUPDWLRQ FRQFHUQLQJ DQQRWDWLRQV SHU GRPDLQ ,W PXVW EH QRWLFHG WKDW DOO DQQRWDWLRQVFRQWDLQWR³VLJQLILFDQW´WHUPVLH QRQVWRSZRUGVWKDW PD\GULYHWKH FRPSXWDWLRQ RI WKH LQWHQGHG PDSSLQJV ,Q DGGLWLRQ WR WKH DERYH PRWLYDWHG E\ RXU H[SHULHQFH ZLWK YHFWRUPRGHOEDVHG RQWRORJ\ DOLJQPHQW PHWKRGV SHUIRUP EHWWHU ZLWKRQWRORJLHVWKDWKDYHULFKLQIRUPDWLRQLHZLWKRQWRORJLHVWKDWFRQWDLQODEHOVDQG FRPPHQWV IRU HYHU\ RQWRORJ\ FODVV  ZH DUWLILFLDOO\ HQULFKHG WKH DQQRWDWLRQ LQIRUPDWLRQRIRQWRORJ\FODVVHVZLWKWH[WXDOLQIRUPDWLRQ$OOH[SHULPHQWV ³HQULFKHG RQWRORJLHV´FDVHV KDYHEHHQUXQZLWKWKHVH³HQULFKHG´RQWRORJLHV  7DEOH,QIRUPDWLRQFRQFHUQLQJDQQRWDWLRQVSHUGRPDLQ  'RPDLQ,G     

$QQRWDWLRQFDVH 'HVFULSWLRQVFDVH E &RPPHQWVFDVH D  'HVFULSWLRQVFDVH E &RPPHQWVFDVH D  'HVFULSWLRQVFDVH E &RPPHQWVFDVH D  'HVFULSWLRQVFDVH E &RPPHQWVFDVH D  'HVFULSWLRQVFDVH F &RPPHQWVFDVH D 

7HUPVWKDWPDWFKDFODVVQDPH $SSUR[ 1$ $SSUR[ $SSUR[ $SSUR[

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¶FRQILJXUDWLRQV)LJXUHVKRZV WKDW DV LW ZDV H[SHFWHG  WKH KLJKHU UHFDOO LV DFKLHYHG E\ WKH FRPELQDWLRQ RI DOO PHWKRGV7KHUHFDOORIDQLQGLYLGXDOPHWKRGIRUDVSHFLILFGRPDLQRQWRORJ\SDLUPD\ QRWEHKLJKGXHWRWKHFKDUDFWHULVWLFVRIWKHVSHFLILFGRPDLQRQWRORJ\SDLU HJGXHWR WKH FRPSRXQG WHUPV LQ QDPH YDOXHV RI WKH :6'/ SDUW HOHPHQWV  )RU LQVWDQFH WKH UHFDOOIRUWKH'RPDLQKRVSLWDO3K\VLFLDQRZOSDLULVORZHYHQIRUFRPELQHGPHWKRGV

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

29

HJ /6$&2&/8  6R HYHQ LI WKH FRPSRVLWLRQ RI WKH ³FRUH PHWKRGV´ RI 86'6 /6$&2&/8  DFKLHYHV UHFDOO HTXDO WR  IRU DOO WKH GRPDLQV LQ DYHUDJH  LW DFKLHYHVIRUWKLVVSHFLILFGRPDLQRQWRORJ\SDLU





)LJ  3UHFLVLRQ DQG UHFDOO IRU GLIIHUHQW 86'6 FRQILJXUDWLRQV XVLQJ DQQRWDWHG :6'/ VSHFLILFDWLRQVDQGHQULFKHGYHUVLRQVRIRQWRORJLHV



)LJ$YHUDJHSUHFLVLRQDQGUHFDOOIRUHDFKFRQILJXUDWLRQ



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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

30

VHOHFWLRQRIWKHPRVW³SURSHU´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¶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





)LJ 3UHFLVLRQ DQG 5HFDOO IRU DOO FRQILJXUDWLRQV LQ H[SHULPHQW 'RPDLQWUDYHORZO ZLWK DQ DUWLILFLDOO\HQULFKHGYHUVLRQRIWKHRQWRORJ\ 

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³TD][VZ´FDQQRWEHGLUHFWO\PDSSHGWRDQ2:/FODVV(YHQWKH/6$PHWKRGWKDW UHOLHVRQ:RUG1HWFDQQRWSHUIRUPZHOO7KLVLVGXHWRWKHIDFWWKDWSDUWQDPHVGRQRW

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

31

KDYH:RUG1HWHQWULHV,QFRQWUDVWWRWKLVWKH960EDVHGPDSSLQJPHWKRGWKDWGRHV QRWUHO\RQDQH[WHUQDOOH[LFRQFDQRYHUFRPHWKLVSUREOHPUHVXOWLQJLQDPLOGO\JRRG SHUIRUPDQFH 7KLV LV VKRZQ LQ WKH H[SHULPHQW FRQILJXUDWLRQ 'RPDLQ%RRNVRZO GHSLFWHG LQ )LJ  7KLV VKRZV WKDW WKH DGHTXDWH DQG ULFK DQQRWDWLRQ RI :6'/ VSHFLILFDWLRQV LV YLWDO WR WKH HIIHFWLYH SHUIRUPDQFH RI WKH PRVW DGHTXDWH PDSSLQJ PHWKRGV7KLVLVIXUWKHUHYLGHQFHGE\WKHUHVXOWVSURYLGHGE\960LQWKHGRPDLQ ZKHUH WKH ³FRPSOH[LW\´ RI WKH WHUPV LQYROYHG SURYLGHV PDMRU REVWDFOHV WR WKH RWKHU PHWKRGV $V D ILQDO FRQFOXVLRQ RI WKH DERYH UHVXOWV ZH FDQ VWDWH WKDW WKH 86'6 FRQILJXUDWLRQ WKDW UHVXOWV IURP WKH FRPSRVLWLRQ RI GLIIHUHQW FRPSOLPHQWDU\ PHWKRGV VHHPVWREHWKHPRVWSURPLVLQJRQH



)LJ  3UHFLVLRQ DQG 5HFDOO IRU H[SHULPHQWV D  'RPDLQERRNVRZO DQG E  'RPDLQ+RVSLWDO3K\VLFLDQRZOZLWKUDQGRPVWULQJV HJ³TD][VZ´ DVSDUWQDPHV

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¶ PDWFKPDNLQJ V\VWHPV HJ >@  7KH WRSN RU WRS.  HYDOXDWLRQ PHWKRG PHDVXUHV WKH SHUFHQWDJH RI FRUUHFW UHVXOWV LQ WKH WRS N UHVXOWV LQ D OLVW RI UDQNHG UHVXOWV DFFXUDF\ RI UHWXUQHG UDQNHG UHVXOWV  ,Q RXU FDVH :6'/ SDUW QDPHV DUH PDWFKHG DJDLQVW RQWRORJ\ FODVVHV DQG WKH PRVW UHOHYDQW SDLUV DUH UHWXUQHG DV VXJJHVWHG PDSSLQJVWRJHWKHUZLWKWKHLUUDQNLQJ7KHKLJKHUUDQNHGSDLUVRIHDFKSDLUDUHEHLQJ DVVHVVHG WR SURYLGH WKH ³FRUUHFW´ PDSSLQJV ,Q FDVH RI FRQIOLFW WKH PHWKRG WKDW SURYLGHG WKH ILUVW PDSSLQJ ZLQV 2I FRXUVH WKLV LV QRW D SURSHU SROLF\ 5DQNLQJV RI GLIIHUHQW PHWKRGV QHHG WR EH QRUPDOL]HG DSSURSULDWHO\ VR DV WR ULFK D JOREDO OLVW RI

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

32

UDQNHG SDLUV +RZHYHU HYHQ LQ WKLV FDVH D VRSKLVWLFDWHG SROLF\ IRU UHFRQFLOLQJ FRQIOLFWV LV QHFHVVDU\ 7KLV LV VRPHWKLQJ WR EH IXUWKHU LQYHVWLJDWHG 7KHUHIRUH WRSN DFFXUDF\ HYDOXDWLRQPHWKRGLGHQWLILHVLIWKH³FRUUHFW´SDLU LHWKHRQHSURYLGHGE\ WKHJROGPDSSLQJV LVDPRQJWKHNUDQNHGVXJJHVWHGPDSSLQJV$FFXUDF\LVDZLGHO\ DFFHSWHG PHWULF LQ WKH ,QIRUPDWLRQ 5HWULHYDO FRPPXQLW\ DQG KDV EHHQ VXFFHVVIXOO\ XVHGIRUWKHHYDOXDWLRQRIPDQ\V\VWHPV$QH[DPSOHRIUHODWHGZRUNFRQFHUQLQJWKH VXSSRUW RI KXPDQ DQQRWDWRUV LV UHSRUWHG LQ >@ 7KH ZRUN LV PRUH IRFXVHG RQ WKH VLPSOLILFDWLRQRIWKHPDQXDODQQRWDWLRQWDVNRIVHUYLFHVUDWKHUWKDQRQWKHYDOLGDWLRQ RIWRSNUDQNHGVHPDQWLFDQQRWDWLRQV)XUWKHUPRUHWKHZRUNIDFLOLWDWHVWKHGHWHFWLQJ RIPLVWDNHVLQH[LVWLQJDQQRWDWLRQVGLVFRYHULQJKRZHYHUDUHODWLYHO\VPDOOQXPEHURI DQQRWDWLRQV 





)LJ7RSSUHFLVLRQIRU960&2&/8/6$/HYHQFRQILJXUDWLRQ 

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

&RQFOXGLQJ5HPDUNV $XWRPDWLF VHUYLFH UHJLVWUDWLRQ WR IDFLOLWDWH ZHE VHUYLFHV¶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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Uncovering WSDL Specifications' Data Semantics

VLQFHLWPDQDJHVWRUHWXUQZLWKKLJKDFFXUDF\WKH³SURSHU´FODVVIRUDQQRWDWLQJHDFK RIWKH:6'/PHVVDJHSDUWVDPRQJWKHKLJKO\UDQNHGPDWFKLQJSDLUV ,Q WKLV SDSHU ZH KDYH SUHVHQWHG H[WHQGHG H[SHULPHQWDWLRQ FDVHV ZLWK GLIIHUHQW FRQILJXUDWLRQVRIWKH86'6PHWKRGIRUXQFRYHULQJZHEVHUYLFHV¶GDWDVHPDQWLFV:H KDYH UHDFKHG D FRQILJXUDWLRQ ZKHUH GLIIHUHQW FRPSOLPHQWDU\ PHWKRGV DFKLHYH YHU\ JRRGSHUIRUPDQFHLVDVHWRIH[SHULPHQWVRIYDU\LQJGLIILFXOW\)XWXUHJRDOVWKDWKDYH EHHQLQGLFDWHGWKURXJKRXWWKHVHFWLRQVRIWKLVDUWLFOHSURYLGHWKHPDQ\GLIIHUHQWIDFHWV RIIXWXUHZRUNWKDWWKLVZRUNHQWDLOV

5HIHUHQFHV  2XQGKDNDU366KHWK$9HUPD.0(7(256:HE6HUYLFH$QQRWDWLRQ)UDPHZRUN :::    6WURXOLD ( :DQJ @:6'/ :HE6HUYLFH'HVFULSWLRQ/DQJXDJH LVDQ ;0/EDVHGODQJXDJHIRUWKHVSHFLILFDWLRQRIZHEVHUYLFHV:6'/LVPDLQO\IRFXVLQJ RQ RSHUDWLRQDO DQG V\QWDFWLF GHWDLOV UHJDUGLQJ WKH LPSOHPHQWDWLRQ DQG H[HFXWLRQ RI :HEVHUYLFHV7KHODFNRIH[SOLFLWVHPDQWLFVLQ:6'/PDNHVVHUYLFHV¶VSHFLILFDWLRQV LQVXIILFLHQWWRVDWLVI\WKHUHTXLUHPHQWVIRUHIIHFWLYH:HEVHUYLFHµPDQLSXODWLRQ¶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¶VHPDQWLF GHVFULSWLRQ 7KH NH\ WHFKQRORJ\ WR WKH ³VHPDQWLF GLVFRYHU\´ DQG WKXV WR WKH ³VHPDQWLF PDWFKPDNLQJ´  RI :HE VHUYLFHV LV RQWRORJLHV 7KH NH\ LGHD KHUH LV WKDW WKH XVH RI   

7KLVZRUNLVVXSSRUWHGSDUWLDOO\E\WKH(XURSHDQ&RPPLVVLRQSURMHFW*ULG$OOXQGHUJUDQW QXPEHU)3SURMHFW,67*ULG$OO

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

100

Look Ma, No Hands: Supporting the semantic discovery of services without ontologies

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³IHWFKHG´ IURP WKH VHUYLFH FRGH RU WKH\ PD\ EH JLYHQ E\ WKH VHUYLFH SURYLGHUV GXULQJ VHUYLFH DGYHUWLVHPHQW 7KLV LQIRUPDWLRQ LV EHLQJ H[SORLWHG IRU WKH FRPSXWDWLRQ RI WKH SRVLWLRQ WKDW :6'/ VSHFLILFDWLRQV FDQ REWDLQ LQ D VHPDQWLF VSDFH ZKHUH WKH GDWD VHPDQWLFV RI :6'/ VSHFLILFDWLRQVDUHH[SUHVVHGE\PHDQVRIDQXPEHURIODWHQWIHDWXUHV*LYHQDTXHU\ WKLVLVFRQVLGHUHGWREHDVSHFLILFDWLRQIRUWKHUHTXHVWHGVHUYLFH¶VGDWDVHPDQWLFV WKH REMHFWLYHRIWKHSURSRVHGPHWKRGLVWRUHSUHVHQWWKLVTXHU\XVLQJWKHFRPSXWHGODWHQW IHDWXUHVDQGILQGWKHUHJLVWHUHG:6'/VSHFLILFDWLRQVWKDWDUHFORVHHQRXJK±DQGWKXV VLPLODU±WRWKHTXHU\LQWKHVHPDQWLFVSDFH'RLQJVRRQHPD\FRQVLGHUWKDWWKHVH ODWHQW IHDWXUHVDUHWKHFRQFHSWVRIDQRQWRORJ\ ZKLFK VHUYHDVLQWHUPHGLDWHV IRUWKH PDWFKLQJRI:6'/VSHFLILFDWLRQV $WWKLVSRLQWZHQHHGWRHPSKDVL]HWKDWDOWKRXJKODWHQWIHDWXUHVDUHEHLQJXVHGIRU H[SUHVVLQJ WKH GDWD VHPDQWLFV RI UHTXHVWHG DQG DGYHUWLVHG VHUYLFHV WKH ODFN RI RQWRORJLHV SURKLELW ORJLFEDVHG LQIHUHQFHV OHDGLQJ WR WKH LQDELOLW\ RI WKH PHWKRG WR H[SOLFLWO\LGHQWLI\LQIHUUHGVXEVXPSWLRQUHODWLRQVEHWZHHQVHUYLFHV¶VSHFLILFDWLRQV 7KH UHVW RI WKH SDSHU LV VWUXFWXUHG LQ WKH IROORZLQJ ZD\ 6HFWLRQ  SURYLGHV WKH UHODWHG ZRUN 6HFWLRQ  SURYLGHV D GHVFULSWLRQ RI WKH SUREOHP DQG WKH EDFNJURXQG WHFKQRORJ\ DQG RXWOLQHV WKH PDWFKPDNLQJ PHWKRG 6HFWLRQ  SUHVHQWV WKH H[SHULPHQWVFRQGXFWHGWRZDUGVHYDOXDWLQJWKHSURSRVHGPHWKRGDQGILQDOO\VHFWLRQ FRQFOXGHVWKHSDSHU

5HODWHG:RUN $OWKRXJK VHYHUDO DSSURDFKHV LQWURGXFH WKH VHPDQWLF PDWFKPDNLQJ RI ZHE VHUYLFHV HJ>DQG@ PRVWRIWKHPUHTXLUHWKHXVHRIDVKDUHGDQGZHOODJUHHG RQWRORJ\ $OWKRXJK WKLV FRQVWUDLQW PD\ EH UHOD[HG E\ H[SORLWLQJ RQWRORJ\ PDSSLQJ WHFKQLTXHV WKHVH WHFKQLTXHV QHHG WR EH IXUWKHU HQKDQFHG WRZDUGV WKHLU JHQHULF GHSOR\PHQW )XUWKHUPRUH WKH GLIILFXOWLHV PHQWLRQHG LQ 6HFWLRQ  UHJDUGLQJ WKH VHPDQWLFGHVFULSWLRQRIZHEVHUYLFHVQHHGWREHWDFNOHG

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Look Ma, No Hands: Supporting the semantic discovery of services without ontologies

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³EDFN´ ZH GHDO ZLWK :6'/ VSHFLILFDWLRQV RI VHUYLFHV VLJQDWXUHV UDWKHU WKDQ ZLWK WKHLU VHPDQWLF FRXQWHUSDUWV LQ 2:/6 DLPLQJ WR VXSSRUW WKH FRQWHQWEDVHG GLVFRYHU\RIVXFKVSHFLILFDWLRQVWRDIXOOH[WHQGUDWKHUWKDQLQFRPELQDWLRQWRORJLF EDVHG DSSURDFKHV 7KLV LV D UDWKHU KDUG SUREOHP JLYHQ WKH VFDUFHQHVV RI WKH LQIRUPDWLRQWKHGHSHQGHQF\RIVSHFLILFDWLRQVRQWKHGHYHORSHUV¶ZKLPDQGWKHVPDOO VL]H RI WKH WH[WXDO GHVFULSWLRQVFRPPHQWV ,Q DGGLWLRQ ZH DLP DW UHWULHYLQJ VSHFLILFDWLRQVWKDWPDWFKH[DFWO\WRDTXHU\ UDWKHUWKDQUHWULHYLQJVSHFLILFDWLRQVWKDW DUHPHUHO\³QHLJKERUV´WRWKHTXHU\  'HDOLQJDOVRZLWK:6'/VSHFLILFDWLRQVDLPLQJWRDGGUHVVWKHFKDOOHQJHVLQYROYHG LQVHDUFKLQJ IRU ZHEVHUYLFHVDXWKRUVLQ>@SUHVHQW:RRJOH D ZHEVHUYLFHVHDUFK HQJLQH,QDGGLWLRQWRVLPSOHNH\ZRUGVHDUFKHV:RRJOHVXSSRUWVVLPLODULW\VHDUFKIRU ZHEVHUYLFHV7KHNH\LQJUHGLHQWRIWKHDSSURDFKLVDUHILQHPHQWRIWKHDJJORPHUDWLYH FOXVWHULQJ RI SDUDPHWHUV¶ WHUPV LQ WKH FROOHFWLRQ RI ZHE VHUYLFHV LQWR VHPDQWLFDOO\ PHDQLQJIXO FRQFHSWV %\ FRPSDULQJ WKH UHVXOWLQJ FRQFHSWV WKLV ZRUN UHSRUWV JRRG VLPLODULW\ PHDVXUHV,QFRQWUDVWWRWKLVDSSURDFKUDWKHUWKDQH[SORLWLQJSDUDPHWHUV¶ WHUPV FRRFFXUUHQFHV ZH DLP DW H[SORLWLQJ WKH KXPDQLQWHQGHG PHDQLQJ RI :6'/ LQSXWRXWSXW PHVVDJHV¶ SDUW HOHPHQWV ³GHVFULELQJ´ WKHP XVLQJ D QXPEHU RI ODWHQW IHDWXUHVDQGSRVLWLRQLQJWKHPLQDODWHQWVSDFH7KLVDSSURDFKLQFRPELQDWLRQWRWKH H[SORLWDWLRQ RI VHUYLFHV¶ WH[WXDO FRPPHQWVGHVFULSWLRQV DV RXU H[SHULPHQWV VKRZ SURYHVWREHYHU\SUHFLVH $QRWKHU ZHEVHUYLFH VHDUFK HQJLQH FRPELQLQJ IRONVRQRPLHV DQG 6HPDQWLF :HE 6HUYLFHVWHFKQRORJLHV >@SUHVHQWVDPHWKRGIRUVHPDQWLFLQGH[LQJDQGDSSUR[LPDWH UHWULHYDO RI :HE VHUYLFHV ,W UHOLHV RQ JUDSKEDVHG LQGH[LQJ LQ ZKLFK FRQQHFWHG VHUYLFHV FDQ EH DSSUR[LPDWHO\ FRPSRVHG ZKLOH JUDSK GLVWDQFH UHSUHVHQWV VHUYLFH UHOHYDQFH $ TXHU\ LQWHUIDFH WUDQVODWHV D XVHU¶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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

101

102

Look Ma, No Hands: Supporting the semantic discovery of services without ontologies

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³PHDQLQJ´E\PHDQVRIODWHQWIHDWXUHVLQGHSHQGHQWO\RIDQ\ H[WHUQDOUHVRXUFH7KHQPDWFKHVEHWZHHQVHUYLFHVLJQDWXUHVDUHGHWHUPLQHGEDVHGRQ VHPDQWLFPDWFKHVRI:6'/LQSXWRXWSXWPHVVDJHVDQGSDUDPHWHUV

6HUYLFH'LVFRYHU\ZLWK/DWHQW6HPDQWLFV 3UREOHPVSHFLILFDWLRQ 7KLVSDSHUGHDOVZLWKWKHIROORZLQJSUREOHP ³*LYHQ D  D UHSRVLWRU\ 5 RI :6'/ VSHFLILFDWLRQV DQG E  D VSHFLILFDWLRQ RI D TXHU\4WKDWVSHFLILHVWKHVLJQDWXUHRIWKHUHTXHVWHGVHUYLFHSURYLGHWKRVHVHUYLFHVLQ 5WKDWPDWFKZLWK4´ $VDOUHDG\SRLQWHGZHGHDORQO\ZLWKVHUYLFHV¶VLJQDWXUHVVSHFLILHGLQ:6'/ LH ZLWKGDWDVHPDQWLFV 0RUHVSHFLILFDOO\  7KH VLJQDWXUH RI D VHUYLFH V VSHFLILHV D VHW RI LQSXWRXWSXW PHVVDJHV 7KHVH DUH GHQRWHG V;L! ZKHUH V LV WKH VHUYLFH LG ; LV WKH W\SH RI WKH PHVVDJH LQSXWRURXWSXWDQGLLVWKHLGRIWKHVSHFLILFPHVVDJH  (DFK PHVVDJH V;L! LV DVVRFLDWHG ZLWK WH[WXDO DQQRWDWLRQV VL@ LQWR RXU IUDPHZRUN ZH GR QRW FRPPLWWRWKHXVHRI6$:6'/DWWKLVVWDJHHPSKDVL]LQJPRVWO\RQWKHXVHRIWH[WXDO GHVFULSWLRQVFRPPHQWVIRU:6'/HOHPHQWV

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Look Ma, No Hands: Supporting the semantic discovery of services without ontologies

,QSXW2XWSXW0HVVDJHV¶VLPLODULW\ /HW XV FRQVLGHU D PHVVDJH V;L! ZKHUH V LV WKH VHUYLFH LG ; LV WKH W\SH RI WKH PHVVDJH LQSXW RU RXWSXW  DQG L LV WKH LG RI WKH VSHFLILF PHVVDJH DV ZHOO DV LWV DVVRFLDWHGWH[WXDODQQRWDWLRQVVL@KDVUHFHQWO\EHFRPHWKH SUHYDOHQWZD\RIEXLOGLQJHQWHUSULVH,QIRUPDWLRQ6\VWHPV ,6 7KHPDLQLGHDEHKLQG 62$LVWKHDELOLW\WRXVHUHXVHDQGVKDUHVHUYLFHVIURPGLIIHUHQWVRXUFHV1RZDGD\V :HE6HUYLFHV :6 DUHWKHSUHYDLOLQJSDUDGLJPIRULPSOHPHQWLQJ62$VXSSRUWHGE\ VLJQLILFDQWLQGXVWU\LQYHVWPHQW>@ ,%0¶V LQLWLDO UHIHUHQFH DUFKLWHFWXUH IRU 62$ >@ LGHQWLILHG WKUHH EDVLF SOD\HUV VHUYLFHSURYLGHUVVHUYLFHUHTXHVWRUVDQGVHUYLFHEURNHUV 7KLVUHIHUHQFHDUFKLWHFWXUH VXSSRUWVIRXUNH\IXQFWLRQDOLWLHVGLVFRYHULQJFRPSRVLQJSXEOLVKLQJDQGLQYRNLQJD VHUYLFH$FRPPRQXVHFDVHLVZKHUHWKHVHUYLFHSURYLGHUSXEOLVKHVVHUYLFHVXVLQJWKH

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

116

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

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³YHUVLRQV´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³DQDUFKLF´ DQQRWDWLRQUHVRXUFH GHVFULSWLRQ SUDFWLFH RI :HE  WR WKH 6HPDQWLF :HE ZRUOG DUH DOUHDG\ XQGHUZD\ ZLWK HIIRUWV VXFKDVWKH0HDQLQJRID7DJ 02$7 SURMHFW>@WKDWKHOSLQDVVLJQLQJPDFKLQH UHDGDEOHPHDQLQJ QDPHO\85,VSRVVLEO\GHILQHGLQDQRQWRORJ\ WRWDJV 7KLVSDSHULVDILUVWDWWHPSWWRPL[H[LVWLQJ6:6GHVFULSWLRQDQGGLVFRYHU\PRGHOV ZLWKWKH:HEWDJDQGXVHUFHQWULFFROODERUDWLYHVROXWLRQVWRUHVRXUFHGLVFRYHU\

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

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



7KH6HUYLFH'LVFRYHU\*DS

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¶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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

117

118

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

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³GHVFULSWLRQ ODQJXDJH´ LH DQ H[SUHVVLRQ ODQJXDJH IRU 2:/6 SURFHVV UHVXOW¶V SUH DQG SRVW FRQGLWLRQVDQGHIIHFWV ,Q VXPPDU\ ZH KDYH REVHUYHG WKDW ZLWKLQ 6:6 GLVFRYHU\ H[DFW ORJLFDO PDWFKPDNLQJLVEHLQJVXSHUVHGHGE\VLPLODULW\EDVHGPDWFKPDNLQJEXWVWLOODWDOHYHO RI ORJLFDO GHVFULSWLRQV 6RPH DSSURDFKHV SURSRVH PXOWLVWHS ILOWHULQJRU VHOHFWLRQ ZKHUH VLPSOH NH\ZRUG PDWFKLQJ RU ,5 PHWKRGV PD\ SUHFHGH ORJLFDO PDWFKLQJ +RZHYHU DOWKRXJK >@ DOUHDG\ HQYLVLRQV ³ZD\ IRU UHTXHVWHUV WR HDVLO\ ORFDWH SUH GH¿QHG JRDOV HJ NH\ZRUG PDWFKLQJ´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µ6HPDQWLF*DS¶>@,QWKHFRQWH[WRIWKLVSDSHUZHPD\WHUPWKLV SUREOHP WKH 6HUYLFH 'LVFRYHU\ *DS  FDXVHG E\ WKH EUHDNGRZQ EHWZHHQ XVHU YRFDEXODU\DQGH[SHFWDWLRQVDQGVHUYLFHGHVFULSWLRQ)RUH[DPSOHWKHXVHU¶VPHQWDO PRGHO PD\ VWUHVV WKH XVHIXOQHVV RI WKH VHUYLFH RXWSXWV LQ WHUPV RI WKH LQSXWV WR D FRPPRQWDVNZKHUHDVWKHVLPLODULW\IXQFWLRQPD\SULPDULO\XVHWKHVHUYLFHLQSXWDQG RSHUDWLRQ IHDWXUHV 7KLV PLVPDWFK PHDQV WKDW WKH XVHU PD\ QRW KDYH D VXLWDEOH YRFDEXODU\WRIRUPXODWHTXHULHVWRUHWULHYHUHOHYDQWVHUYLFHV 7KLV JDS KDV PDQ\ PRUH IDFHV WKDQ RQO\ WKH GLVFUHSDQF\ EHWZHHQ WHFKQLFDO GHVFULSWLRQVRIVHUYLFHVDQGWKHYRFDEXODU\WKDWPLJKWEHXVHGIRUNH\ZRUGVHDUFKE\ SRWHQWLDOUHTXHVWHUV  /DQJXDJH*DSWKHUHLVDODFNRIFRPPRQVHPDQWLFGHVFULSWLRQVRIDYDLODEOHVHUYLFHV E\WKHLUSURYLGHUV:KHUHDYDLODEOHDWDOOVHUYLFHGHVFULSWLRQVPD\XVHGHVFULSWLRQVLQ YDU\LQJIRUPDWVOHYHOVRIJUDQXODULW\DQGXQGHUO\LQJ ORJLFDO ODQJXDJHV'HVSLWHWKLV PRVW VHUYLFH IUDPHZRUNV FRQWDLQ WKH VDPH VHUYLFH DVSHFWV FRQWDLQHG LQ :6'/ ,QSXWV 2XWSXWV 2SHUDWLRQV DQG PRUH UDUHO\ 3UHFRQGLWLRQV DQG (IIHFWV ZKLFK DUH IRULQVWDQFHPLVVLQJLQ6$:6'/   3URYLGHU8VHU *DS 7KLV JDS RFFXUV ZKHUH SURYLGHUV KDYH GLIIHUHQW LQWHQWLRQV IRU WKHXVHRIWKHLUVHUYLFHWKDQWKHXVHUVZKRFRQVXPHWKHVHUYLFH

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV



:HE$SSURDFKHVWR6HUYLFH'LVFRYHU\

7RDGGUHVVWKLVSUREOHPZHORRNDWKRZXVHUGHILQHGWDJVPLJKWKHOS7DJVDUHVKRUW LQIRUPDO GHVFULSWLRQV RIWHQ RQH RU WZR ZRUGV ORQJ XVHG E\ :HE XVHUV WR GHVFULEH RQOLQH UHVRXUFHV 7KHUH DUH QR WHFKQLTXHV IRU VSHFLI\LQJ ³PHDQLQJ´ RU LQIHUULQJ RU GHVFULELQJUHODWLRQVKLSVEHWZHHQWDJV7DJFORXGVUHIHUWRDJJUHJDWHGWDJLQIRUPDWLRQ LQ ZKLFK D WD[RQRP\ RU ³WDJVRQRP\´ HPHUJHV WKURXJK UHSHDWHG FROOHFWLYH XVDJH RI WKHVDPHWDJV3DUW$RI)LJXUHLOOXVWUDWHVDWDJFORXGLQWKHEORJGRPDLQ 7KHDGYDQWDJHRIXVLQJWDJVLQWKHFRQWH[WRIVHUYLFHGLVFRYHU\LVWKDWWKH\VXSSO\D XVHUGHILQHGYRFDEXODU\EDVHGRQDFRQVHQVXVRIKRZWKHVHUYLFHLVSHUFHLYHGRUXVHG LQWKHZRUOG)RUWKHQHHGVRIRXUUHVHDUFKZHDVVXPHWKDWHDFKVHUYLFHWKDWLVPDGH DYDLODEOHYLDWKH:HEKDVLWVRZQWDJFORXG

 )LJXUH$JOREDOWDJFORXGDQGVHYHUDOORFDOµV\QWKHWLF¶WDJFORXGVZKLFKZHUHSURGXFHGE\ FOXVWHULQJWKHDVVRFLDWHGEORJFRQWHQW>@



0$&)$&6LPLODULW\0DWFKLQJ

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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

119

120

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

>@7KHHVVHQWLDOREVHUYDWLRQLVWKDWUHWULHYDOFDQWDNHSODFHLQFUHPHQWDOO\LQDWOHDVW WZRVWHSVZKHUHWKHLQLWLDOVWHSVLQYROYHWKHUHWULHYDORIDVXEVHWRILWHPE\PHDQVRI DQ LQH[SHQVLYH IXQFWLRQ RQ VXUIDFHOHYHO IHDWXUHV IROORZHG E\ D UHILQHPHQW RI WKLV VXEVHWXVLQJPRUHVRSKLVWLFDWHGWHFKQLTXHVRQVWUXFWXUDOIHDWXUHV  $SSURDFK0DWFKLQJ7DJFORXGV  7KLVPRGHODOORZVXVWRLQFRUSRUDWHWDJLQIRUPDWLRQDQGIRUPDOVHUYLFHGHVFULSWLRQV LQWRDXVHUFHQWULFVHUYLFHGLVFRYHU\PRGHO,IZHWDNHDVWHSEDFNDQGUHWKLQNZKLFK ³VHPDQWLFGHVFULSWLRQV´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³ILOWHU´WR PDVNHDFKVHUYLFHWDJFORXGGURSSLQJDOOWDJVWKDWDUHQRWRILQWHUHVWDQGWKHZHLJKWHG VXPRIWKLVPDVNHGWDJFORXGZRXOGGHQRWHWKHGHJUHHRIPDWFK 7R FRPSDUH WDJFORXGV ZHLJKWV DUH XVXDOO\ QRUPDOLVHG 7KH QRUPDOLVHG WDJ IUHTXHQF\ULRIWDJWLLQ&ZKHUHNLVWKHQXPEHURIWDJVLQ&LVGHILQHGDV QL  UL = ¦ QN N

/HW7EHDQRUPDOLVHGWDJFORXGDQG4DXVHUVSHFLILHGWDJVHWWKHQWKHVLPLODULW\ EHWZHHQ7DQG4LVGHILQHGDV  VLP 7  4 = ¦ δ WSL  4  WL ∈7

ZKHUHWSLVWDQGVIRUWKHQRUPDOLVHGWDJSDLU WDJZHLJKW! DQGδLVGHILQHGDV  ­U LI W ∈ 4  δ < W  U > 4 = ®  ¯  RWKHUZLVH  7KLVQDwYHDSSURDFKRIWDJFORXGPDWFKLQJFRXOGREYLRXVO\EHUHILQHGE\OHVVRQV OHDUQHGIURPWKHWUDGLWLRQDO VHPDQWLFVHUYLFH PDWFKPDNLQJUHDOP7\SLFDOSURSRVDOV WRVHUYLFHPDWFKLQJXVHDV\PPHWULFPHDVXUHVLQWKHGHJUHHRIPDWFKEHWZHHQUHTXHVW DQGRIIHUZKLFKWDNHVLQWRDFFRXQWWKHVXEVXPSWLRQUHODWLRQEHWZHHQWKHP HJSOXJ LQ YV VXEVXPHV LQ 3DROXFFL¶V >@ DSSURDFK  )RU H[DPSOH VXSSRVH WZR VHUYLFH

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

GHVFULSWLRQVV EX\ERRNDQGV EX\ILFWLRQERRNWKHGHJUHHRIPDWFKLVXVXDOO\ GLIIHUHQW LI WKH V LV WKH XVHU UHTXHVW DQG V LV WKH SURYLGHU RU YLFH YHUVD 7KLV DV\PPHWU\ LV ORVW LQ WKH WDJ FORXG WKHUH LV QR VXEVXPSWLRQ UHODWLRQ EHWZHHQ FRQFHSWV EXWPLJKWEHHPXODWHGE\XVLQJWKHVXEVHWRSHUDWRUEHWZHHQWDJVHWV HJ ^EX\ERRN`⊆^EX\ILFWLRQERRN`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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

121

122

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

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¶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

¦Q

& S = Q∈1

1



 7KHFOXVWHUFHQWURLGLVDYHFWRUWKDWFRQWDLQVDZHLJKWHGUHSUHVHQWDWLRQRIWKHWDJV PRVWUHSUHVHQWDWLYHRIWKHFRQFHSWLQFOXVWHU$V\QWKHWLFWDJFORXGFDQEHH[WUDFWHG IURPWKHFHQWURLGXVLQJD WKUHVKROGWRILOWHUORZO\ ZHLJKWHGWHUPVDQG XVLQJWKH WDJ WHUPZHLJKWVDVDQLQSXWWRDQ\WDJFORXGSUHVHQWDWLRQDOJRULWKP

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

7KHRYHUDOORXWSXWLVDEURZVHDEOHGHQGRJUDPZKHUHHDFKQRGHLVUHSUHVHQWHGE\ DV\QWKHWLFWDJFORXGUHSUHVHQWLQJVHUYLFHGHVFULSWLRQDWGLIIHUHQWOHYHOVRIVSHFLILFLW\ VHH)LJXUH 

 )LJXUH  $Q LOOXVWUDWLRQ RI D GHQGURJUDP LQGXFHG IURP WDJ FORXG GDWD $W HDFK QRGH D WKUHVKROG FRQWUROVKRZ PDQ\ WDJV DUH GLVSOD\HG )RU HDFK VXEQRGH WKH VLQJOH PRVWO\ KLJKO\ ZHLJKWHGWDJLQHDFKSDUHQWQRGHLVGLVSOD\HGDVDSDUWRIDSDWKVXPPDU\(JERRNV!EX\

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µFRQWHQW¶ ORRVHO\ UHIHUV WR DQ\ QRQVRFLDOO\ GHULYHG GHVFULSWLYH GDWD WKDW FDQ EH XVHG IRU UHWULHYDOSXUSRVHVW\SLFDOO\XVLQJ,5LQVSLUHGPDWFKLQJDOJRULWKPV ,QWKHFRQWH[WRIWKLVZRUNZHSODQWROHYHUDJHDµFRQWHQW¶EDVHGDSSURDFKWRWKH µFROGVWDUW¶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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

123

124

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

DFW DV LQSXW IRU D FROGVWDUW FOXVWHULQJ SURFHVV 7KH FOXVWHU FRQFHSWV ZLOO EH UHSUHVHQWHGE\V\QWKHWLFWDJFORXGVH[WUDFWHGIURPWKHDYDLODEOHVHUYLFHGHVFULSWLRQV SURYLGHUV¶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



&RQFOXVLRQVDQG)XWXUH:RUN

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¶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

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

$FNQRZOHGJPHQWV 7KLVUHVHDUFKKDVEHHQSDUWLDOO\VXSSRUWHGE\WKH(XURSHDQ&RPPLVVLRQXQGHUJUDQW )3,67 LQ&RQWH[W  E\ WKH 6SDQLVK 0LQLVWU\ RI (GXFDWLRQ DQG 6FLHQFH WKURXJK SURMHFWV 85-&&0&(7 DQG &2162/,'(5 &6' ,1*(1,2 $JUHHPHQW7HFKQRORJLHV DVZHOODVE\6FLHQFH)RXQGDWLRQ,UHODQG XQGHUWKH/tRQSURMHFW 6),&(, 

5HIHUHQFHV   

   





   

$NNLUDMX 5 HW DO :HE 6HUYLFH 6HPDQWLFV  :6'/6 :& 0HPEHU 6XEPLVVLRQKWWSZZZZRUJ6XEPLVVLRQ:6'/6 $QNROHNDU$HWDO'$0/66HPDQWLF0DUNXS)RU:HE6HUYLFHV3URFHHGLQJV RIWKH,QWHUQDWLRQDO6HPDQWLF:HE:RUNVKRS   $YHVDQL 3 +D\HV & &RYD 0   /DQJXDJH *DPHV 6ROYLQJ WKH 9RFDEXODU\ 3UREOHP LQ 0XOWL&DVH%DVH 5HDVRQLQJ LQ SURFHHGLQJV RI WKH WK ,QWHUQDWLRQDO &RQIHUHQFH RQ &DVH%DVHG 5HDVRQLQJ ,&&%5  +HFWRU 0XQR] $YLOD)UDQFLVFR5LFFL HGV 6SULQJHU %HUQHUV/HH 7 +HQGOHU - /DVVLOD 2/ 7KH VHPDQWLF ZHE 6FLHQWLILF $PHULFDQYRO  SS±   %HUQVWHLQ$.OHLQ07RZDUGVKLJKSUHFLVLRQVHUYLFHUHWULHYDO,(((,QWHUQHW &RPSXWLQJYRO     %XUNH 5+\EULG5HFRPPHQGHU 6\VWHPV6XUYH\DQG([SHULPHQWV8VHU 0RGHOLQJDQG8VHU$GDSWHG,QWHUDFWLRQ 1RY  &KHQ + 1J 7 ' 0DUWLQH] - DQG 6FKDW] % 5  $ FRQFHSW VSDFH DSSURDFKWRDGGUHVVLQJWKHYRFDEXODU\SUREOHPLQVFLHQWLILFLQIRUPDWLRQUHWULHYDO DQH[SHULPHQWRQWKHZRUPFRPPXQLW\V\VWHP-$P6RF,QI6FL -DQ   G¶$PDWR & 6WDDE 6 )DQL]]L 1 DQG ) (VSRVLWR (IILFLHQW GLVFRYHU\ RI VHUYLFHVVSHFLILHGLQGHVFULSWLRQORJLFVODQJXDJHV,Q3URFHHGLQJVRIWKH,6:& :RUNVKRSRQ6HUYLFH0DWFKPDNLQJDQG5HVRXUFH5HWULHYDOLQWKH6HPDQWLF :HE 605  )HUQDQGH] $ 3ROOHUHV $ DQG 2VVRZVNL 6 7RZDUGV )LQHJUDLQHG 6HUYLFH 0DWFKPDNLQJ E\ 8VLQJ &RQFHSW 6LPLODULW\ ,Q 3URFHHGLQJV RI WKH ,6:& :RUNVKRSRQ6HUYLFH0DWFKPDNLQJDQG5HVRXUFH5HWULHYDOLQWKH6HPDQWLF:HE 605  )RUEXV . ' *HQWQHU ' DQG /DZ .  0$&)$& $ 0RGHO RI 6LPLODULW\EDVHG5HWULHYDO&RJQLWLYH6FLHQFH )XUQDV * : /DQGDXHU 7 . *RPH] / 0 DQG 'XPDLV 6 7  7KH YRFDEXODU\ SUREOHP LQ KXPDQV\VWHP FRPPXQLFDWLRQ &RPPXQ $&0   1RY  *LVROIL ' ,%0 $Q LQWURGXFWLRQ WR G\QDPLF HEXVLQHVV KWWSZZZLEPFRPGHYHORSHUZRUNVZHEVHUYLFHVOLEUDU\ZVDUF   *ULPP 6 0RWLN % 3UHLVW & 9DULDQFH LQ HEXVLQHVV VHUYLFH GLVFRYHU\ ,Q 3URFHHGLQJVRIWKH,6:&:RUNVKRSRQ6HPDQWLF:HE6HUYLFHV  

2nd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web

125

126

Closing the Service Discovery Gap by Collaborative Tagging and Clustering Techniques

&ORVLQJWKH6HUYLFH'LVFRYHU\*DSE\&ROODERUDWLYH7DJJLQJDQG&OXVWHULQJ7HFKQLTXHV

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²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±  /L