Ontologies


While carrying out my research I’ve had to produce a number of ontologies or contribute to existing ones. Find here some of those that I’ve managed to collect quickly. A good deal of these ontologies are in OCML for it is the language used internally by the Internet Reasoning Service. It is a pretty powerful language and quite easy to understand  so don’t be scared by the name. I include in some cases translations into other languages such as OWL if I’ve had the time to do so. Hopefully I will be able to update this section frequently as I get ahold of old work or produce new stuff.

Linked USDL

Linked USDL is an RDF(S) vocabulary that aims to better promote and support the trading of services on the Web. The kinds of services targeted for coverage by USDL include human services (e.g., consultancy), business services (e.g. purchase order requisition), software services (e.g., WSDL and RESTful services), infrastructure services (e.g., CPU and storage services), etc.

Linked USDL is a remodelled version of USDL (an earlier initiative from SAP Research), that builds upon the Linked Data principles and the Web of Data.

See Linked-USDL.org for the latest details and for accessing the ontologies or this page for a short summary.

Linked USDL Core

Linked USDL Core is the foundation module of Linked USDL and covers four essential aspects:

  • Service descriptions
  • Service offering descriptions
  • Business entities involved in the service delivery chain
  • Interaction points allowing consumers to contract or trigger the benefits of contracted services

Linked USDL Core relies on a number of pre-existing vocabularies to supporting this including, DC Terms, SKOS, Time Ontology, vCard, Minimal Service Model (MSM), and GoodRelations.

See Linked USDL Core Web page for further details.

Interaction Roles

In order to capture the roles played by involved entities within particular interactions during the life-cycle of a service, Linked USDL includes a SKOS categorisation with the most typical ones. This categorisation has purposely been kept minimal, is optional and can indeed be replaced and/or extended as necessary. The current Interaction Roles contemplated are as follows: Observer, Participant, Initiator, Mediator, Receiver.

See Interaction Roles Web page for further details.

Business Roles

In order to capture the business roles played by involved entities in a service, Linked USDL includes a SKOS categorisation with the most typical business roles. This categorisation has purposely been kept minimal, is optional and can indeed be replaced and/or extended as necessary. The categorisation includes the following roles: Provider, Producer, Regulator, Intermediary, Consumer, Customer.

See Business Roles Web page for further details.

 

Minimal Service Model

Minimal Service Model (MSM) is a simple RDF(S) ontology for capturing the semantics of remotely invocable software components including notably Web APIs, Web services, and sensors. Thus, MSM does not aim to be yet another service model to bring further heterogeneity to the services landscape. It is instead an integration model at the intersection of existing formalisms and based on the principle of minimal ontological commitment. MSM is a common model able to homogeneously support the publication and discovery over Web APIs, Web services and sensors. This vocabulary is core model used within our Linked Services registry iServe.

MSM has been enriched and integrated with widely used vocabularies for additional coverage including Dublin Core, FOAF, and WSMO-Lite to name a few.

See the github project for the sources and a Jena-based Java library for manipulating MSM descriptions.

See its Web page for documentation.

COBRA

COBRA provides a core terminology for supporting Business Process Analysis where analysts can map knowledge about some particular domain of interest in order to carry out their analyses. It is worth noting that COBRA does not aim to provide a fully-comprehensive conceptualisation for supporting each and every kind of analysis since the scope would simply be too big to be tackled appropriately in one ontology. Instead COBRA provides a pluggable framework based on the core conceptualisations required for supporting Business Process Analysis and defines the appropriate hooks for further extensions in order to cope with the wide-range of aspects involved in analysing business processes. COBRA and its extensions have mostly been developed in the context of SUPER as part of an ontological framework aimed at providing an extensive conceptualisation of the Business Process Management domain ranging from process modelling to the definition of business strategies.

[download id=”28″ format=”1″]

COBRA in OWL (translation by Dong Liu)

Events Ontology

We have extended COBRA with a reference Events Ontology (EVO) that provides a set of definitions suitable for monitoring and a large variety of systems and ready to be integrated within our core ontology for analysing business processes. EVO is based on previously devised formats for monitoring Business Process Management Systems such as MXML, since they provide general purpose solutions that have shown to be suitable to capture logs generated by a plethora of systems. As prescribed by COBRA, EVO is centred around a state model that accounts for the status of processes and activities and it allows to keep track of the evolution of the execution of processes and their internal activities through transitions within the state model. EVO shared many aspects with other provenance models although it has been essentially inspired by previous work on BPM rather than provenance analysis.

[download id=”30″ format=”1″]

Events Analysis Ontology

In order to derive information upon reception of monitoring events we have defined in Events Analysis Ontology, a set of generic forward-chaining rules which are independent from the domain and the specific Monitoring Events defined. The goal is to provide reusable rules which can then be enhanced with domain specific ones to derive a richer knowledge-base. Additionally we have implemented a set of relations which are of most use when analysing processes. Some of these relations have been defined for COBRA in a generic manner, whereas others have been bundled with EVO for they are EVO-specific.

[download id=”29″ format=”1″]

Time Ontology

Time Ontology provides a temporal reference by means of which one can determine temporal relations between entities. The ontology defines three top-level concepts, namely Time Instant, Time Interval, and Temporal Entity. Time Instant is the main primitive element and it provides the means for identifying a point in time with precision up to the microsecond for we aim to support monitoring automated systems. Time Intervals are defined by means of the start and end instants and have therefore an associated duration which can be computed by means of a function that substracts the limiting instants. Temporal Entity, as opposed to the conceptualisation proposed at W3C, represents entities that have a temporal occurrence, and are therefore different from Time Instant and Time Interval which are the base constructs that represent a particular point or period in time. Using these core concepts we have implemented the interval relations defined by James Allen, the additional instant-interval relations defined by Marc Vilain, and useful functions for computing the duration of intervals or for obtaining the current Time Instant.

[download id=”39″ format=”1″]

Metrics Ontology

[download id=”35″ format=”1″]

Generic Business Metrics

[download id=”31″ format=”1″]

Generic Business Queries

[download id=”32″ format=”1″]

Metrics Computation Task

[download id=”34″ format=”1″]

Physical Quantities

[download id=”36″ format=”1″]

Standard Prefixes

[download id=”38″ format=”1″]

Standard Dimensions

[download id=”37″ format=”1″]

Units Manipulation

[download id=”40″ format=”1″]

International System of Units

[download id=”33″ format=”1″]