Views
3 years ago

Semi-Automatically Annotating Semantic Web ... - Andreas Hess

Semi-Automatically Annotating Semantic Web ... - Andreas Hess

Semi-Automatically Annotating Semantic Web ... - Andreas

Semi-Automatically Annotating Semantic Web Services (Extended Abstract)Andreas Heß Eddie Johnston Nicholas KushmerickComputer Science Department, University College Dublin, Ireland{andreas.hess, eddie.johnston, nick}@ucd.ieOverviewThe semantic Web Services vision requires that each servicebe annotated with semantic metadata. Various metadatalanguages (such as OWL-S (DAML-S Coalition 2003))have been proposed to fill this “semantic gap”. However,manually creating such metadata is tedious and error-prone.Software engineers, accustomed to tools that automaticallygenerate WSDL, might not want to invest the required effort.This extended abstract describes ASSAM, a tool that assistsa user in creating semantic metadata for Web Services.ASSAM’s capabilities to automatically create semanticmetadata are supported by two machine learning algorithms.First, we have developed an iterative relational classificationalgorithm for semantically classifying Web Services,their operations, and input and output messages. Second,to aggregate the data returned by multiple semanticallyrelated Web Services, we have developed a schema mappingalgorithm based ensembles of string distance metrics.The remainder of this paper is organized as follows: (1)We describe the WSDL annotator application. This componentof ASSAM uses machine learning to provide the userwith suggestions on how to annotate the elements in theWSDL. (2) We describe and evaluate the iterative relationalclassification algorithm that provides these suggestions. (3)We describe and evaluate OATS, a schema mapping algorithmspecifically designed for the Web Services context.This extended abstract merely summarizes our efforts; fordetails and a discussion of related work, see (Heß & Kushmerick2004; Johnston & Kushmerick 2004; Heß, Johnston,& Kushmerick 2004).ASSAM: A Tool for Web Service AnnotationOne of the central parts of ASSAM is the WSDL annotatorapplication. The WSDL annotator is a tool that enables theuser to semantically annotate a Web Service using a pointand-clickinterface. The key feature of the WSDL annotatoris the ability to suggest which ontological class to use toannotate each element in the WSDL.ASSAM is designed primarily for users who want to annotatemany similar services. These users could be enduserswanting to integrate several similor Web Services intohis or her business processes, or the administrators of a centralizedsemantic Web Service registry. Our tool could alsobe useful for programmers who are only interested in annotatinga single Web Service they have created. In order tomake his or her service compatible with existing services, adeveloper might want to annotate it with the same ontologythat has already been used for some other Web Services. Thedeveloper could import the existing Web Services in AS-SAM and use them as training data in order to obtain recommendationson how to annotate his or her own service.Fig. 1 shows the ASSAM application. Note that our application’skey novelty—the suggested annotations createdautomatically by our machine learning algorithm—is shownin the small pop-up window.The left column in the main window contains a list of WebServices and the category ontology. Web Services can beassociated with a category by clicking on a service in a listand then on a node in the category tree. When the user hasselected a service and wants to focus on annotating it thispart of the window can be hidden.The middle of the window contains a tree view of theWSDL. Port types, operations, messages and complex XMLschema types are parsed from the WSDL and shown in a treestructure. The original WSDL file is also shown as well asplain text descriptions from the occasional documentationtags within the WSDL or a plain text description of the serviceas a whole, such as often offered by a UDDI registry ora Web Service indexing web site.When the user clicks on an element in the WSDL treeview, the corresponding ontology is shown in the right columnand the user can select an appropriate class by clickingon an element in the ontology view. Currently, different ontologiesfor datatypes and operations are used. At presentwe allow annotation for operations, message parts and XMLschema types and their elements. Port types or messagescannot be annotated, because there is no real semantic meaningassociated with the port type or the message itself that isnot covered by the annotation of the operations or the messageparts.Because we do not handle composition and workflow inour machine learning approach, the generated process modelconsists only of one atomic process per operation. The generatedprofile is a subclass from the assigned category of theservice as a whole – the category ontology services as profilehierarchy. The concept file contains a representation ofthe annotated XML schema types in OWL-S. Note that it is

antiRadiant: Tool for Semantic Annotations for Web Service ... - LPIS
Machine Learning for Annotating Semantic Web ... - Andreas Hess
Machine Learning for Annotating Semantic Web ... - Andreas Hess
ASSAM: A Tool for Semi-Automatically Annotating ... - CiteSeerX
Machine learning techniques for annotating semantic ... - CiteSeerX
Learning to attach semantic metadata to Web Services - Andreas Hess
Semantic Annotation for Semi-Automatic Positioning of the ... - Ontotext
A Tool for Semi-Automatically Annotating Semantic Web Services
Supervised and Unsupervised Ensemble Learning ... - Andreas Hess
CONTENTUS - towards semantic multi-media libraries - Andreas Hess
Automatic Image Annotation by Mining the Web - Department of ...
Playful Validation of Automatically Extracted Data - Andreas Hess
Automatic Semantic-Role Annotation for Portuguese - VISL
A Semi-Automatic Semantic Annotation and Authoring Tool for a ...
An Iterative Algorithm for Ontology Mapping ... - Andreas Hess
A Classification of Semantic Annotation Systems - Semantic Web ...
Semantic Annotation, Indexing, and Retrieval - Web Semantics
Iterative Ensemble Classification for Relational Data ... - Andreas Hess
Automatic Semantic Annotation with KIM - 3rd International ...
A Classification of Semantic Annotation Systems - Semantic Web ...
Benchmarking ontology-based annotation tools for the Semantic Web
Annotation of Heterogeneous Database Content for the Semantic Web
Annotation of Heterogeneous Database Content for the Semantic Web
A System for A Semi-Automatic Ontology Annotation - LT4eL
Semi-automatic compound nouns annotation for ... - ResearchGate
Injecting semantic annotations into (geospatial) Web service ...
Injecting semantic annotations into (geospatial) Web service ...
Automatic Annotation for All Semantic Layers in FrameNet
Automatic image annotation using a semi-supervised ensemble of ...