Download presentation
Presentation is loading. Please wait.
Published byAnabel Neal Modified over 8 years ago
1
Lisbon, 30 th March 2016 Gianluca Luraschi gianluca.luraschi@gmail.com Gonçalo Cadete goncalo.cadete@tecnico.ulisboa.pt “Towards a Methodology for Building Ontologies” Mike Uschold and Martin King Ontology Engineering “The Enterprise Ontology” Mike Uschold, Martin King, Stuart Moralee and Yannis Zorgios
2
Presentation 1.Introduce the skeleton methodology for building an Ontology. 2.Apply the methodology to a specific case study. 2
3
A Skeletal Methodology - stages 3 Integrating existing Ontologies Integrating existing Ontologies 3.Evaluation How to build an Ontology 1.Identify Purpose 2.Building the Ontology Ontology capture Ontology coding 4.Documentation What is an Ontology? “an explicit specification of a conceptualization” (by Gruber)
4
Purpose stage 4 scope: 1.Why the Ontology is being built. 2.What its intended uses are. 3.Identify and characterize the users. Case Study – EMSA’s Earth Observation Ontology WHY/WHAT - The main aim of this project is to agree about set of terms and their definitions among all the EMSA’s EO stakeholders in order to avoid ambiguities in using the EO concepts (inter-lingua). WHO – The users of this ontology are the EMSA’s EO stakeholders: EMSA’s EO suppliers (ICT and EO services providers), EO business and technical teams, EMSA’s EO clients (Frontex and EFCA, DG Move).
5
Building the Ontology - Capture Step 5 Scope: 1.Identify the key concepts and relationship in the domain of interest; 2.Production of precise unambiguous text definition for such concepts and relationship; 3.Identification of terms to refer such concepts and relationship; 4.Agreeing on all the above. -> Issue: Intermediate representation of a conceptualization -> more/less formal/structured Approach: brainstorming; grouping -> Case Study (EMSA’s EO Ontology) Extraction of concepts, terms and definition for EODC business requirements, EODC contracts, EO tender specifications and procured Requests For Change. Tips: Initially do not formally commit to any particular meta-ontology. Definition effort/degree agreement. Handling ambiguities. Guidelines to follow. Wording. Reviewing How: “If we view the real world as a set of objects in a universe of discourse then identify aspects of interest can be seen as choosing and defining categories of objects found in the universe” - > issue: granularity of the categories
6
Building the Ontology - Coding 6 Methodology’s stage scope: Explicit representation of the conceptualization captured Issue: Choosing language and a methods -> see the scope of the Ontology Note: sometimes capture and coding are merged into a single step -> Tips: do not do that Case Study – EMSA’s EO Ontology Criteria to be addressed for language: perspicuity (clarity); conceptual distance (how complex is using the primitive to translate the concepts); expressive power (will it represent the concepts we need?); standards compliancy; translability/transportability; support by any methods/guidelines; does it have any formal semantics? (consistency checks); easiness; what kind of user base does it have?; flexibility. Criteria to be addressed for methods: coverage/completeness (how many stages does it cover?); granularity (does it address things at the right level?); learning curve (is it easy to learn?)
7
Building an Ontology - Integrating 7 Methodology’s stage scope: how and whether to use ontologies that already exist Issue: Much work must be done to achieve agreement -> tips: make explicit all assumptions underlying the ontology Case Study – EMSA’s EO Ontology Actions: 1.survey of EO ontology already existing 2.Benchmark the terms extracted and the EO ontology
8
Evaluation - Documentation 8 Evaluation scope: to make a technical judgment of the ontology Documentation scope: to use an ontology Issue: 1. One of the main barriers to effective knowledge sharing is the inadequate documentation -> tips: all the important assumptions shall be documented for the concepts as well as the primitives
9
Findings 9 TermsAcronymsDefinitionDefinition's source Acquisition The tasking of satellite acquisitions is done directly by the CleanSeaNet service with the Service Providers (SPs). The SP’s offer a service chain to EMSA including (i) satellite data acquisition at Ground Station(s), (ii) NRT processing of the SAR data, (iii) value adding image analysis services and (iv) delivery of products and associated metadata to EMSA, which all have to be performed according to satellite overpass at any time of the day. Alerting component Graphical User Interface ALERTING GUI It will be a separate component not part of the SEG. The business requirements should cover the needs for all EO based maritime services e.g. CSN, Frontex. Accordingly, the business requirements are not covered in this document. allocation The tasking of satellite acquisition is done directly by the CleanSeaNet service. The tasking is sent to the operational contact points within the Coastal States for confirmation. This process is called “image allocation”, and is used by Coastal States when planning the aerial and/or vessel verification. When a Coastal State has confirmed the tasking, EMSA issues the Tasking Forms for the scenes to the Satellite Operators and Service Providers. API In computer programming, an application programming interface (API) is a set of routines, protocols, and tools for building software applications. An API expresses a software component in terms of its operations, inputs, outputs, and underlying types Earth Observation Processing – Technical Specification Version 2.0 Next Steps: 1. EO ontology survey; 2. agree among the stakeholders about the definition; 3. handling the ambiguities; 4. review the terms; 5. devise a meta-ontology
10
Conclusion 10 1.Interesting starting point but it looks like very old paper (1995) and it is not providing a systematic approach. 2.based on the exercise the paper’s tips are still valid.
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.