A model for the integration of multiple models.

Metadata

IRI

https://linked.data.gov.au/def/supermodel

Preferred Label

Supermodel

Definition

A Semantic Web model aimed at facilitating dataset-specific and generalised, integrative, modelling within scenarios.

Created

2021-12-10

Modified

2025-02-15

Issued

0000-00-00

Creator

Nicholas J. Car

Publisher

-

This model is not officially published by an Australian government organisation

Provenance

This model has been created in response to the needs of several projects that require both specialised component models and yet integration of them. The projects were: Geoscience Australia’s Sites, Samples Surveys modelling and Linked Data systems upgrade in 2022, FSDF operationalisation 2022, Queensland Spatial Information Addressing & Cadastral modelling 2022 and the Indigenous Data Network catalogue modelling, 2021-2023.

Status

Draft

Version

sup:0.0.1

Code Repository

https://github.com/nicholascar/supermodel

License

Creative Commons Attribution 4.0 International (CC BY 4.0)

Copyright

© Nicholas J. Car, 2021-2025

Machine-readable form (RDF)

https://linked.data.gov.au/def/supermodel.ttl

Introduction

GSWA Supermodel

An informal overview of the GSWA Supermodel and its various parts.

This model provides elements to relate other models to one another and to assign them roles indicating what parts they play with respect to one another. In doing this, this Supermodel enables both the establishment of specialised models for an initiative’s specialised domain data and also for their integration.

Supermodels require that a set of models be nominated for use (they may be newly created but are often taken from pre-existing, fairly generic, models) and given the roles of Integrative Model. All data within the Supermodel scope is expected to be implemented according to elements and patterns from them, at least generically. Elements and patterns from these models are what people can expect to find in use at all points in within the Supermodel, even if parts of it extend beyond these into very specialised versions of them.

Models with role Component Model specialise the integrative models' elements and patterns can be created for each dataset within the Supermodel’s remit. These Component Models can implement the specialised elements and patterns needed to represent their unique content but MUST be fully compatible with the Integrative Models. To ensure this is the case, profiling mechanisms from The Profiles Vocabulary [PROF] are used to relate Component and Integrative models.

The total set of identified existing models from which Component Models draw existing elements and patterns but which are not the common integrative models are given the role Background Model. The Supermodel mechanisms require that all model elements and patterns not created specifically for its domain work come from identified models with either the Integrative or Background model roles.

Supermodel Model

I don’t think modeling is a career I’m going to pursue.
— Kate Moss

Roles

Each model and vocabulary included within an instance of a Supermodel has at least one of the following Roles assigned to it:

Integrative Model

Role for a model that is specialised and extended by multiple models with Component Model role

Component Model

Role for individual datasets within a Supermodels' remit which specialise and extend models with the Integrative and Background Model roles

Background Model

Role for a model that is specialised and extended by some models with Component Model role but not all

Vocabulary

Role for a controlled list of terms used in any of the Supermodel’s models

Mechanisms

The Supermodel and models within it must use a consistent modelling mechanism that allows for the modelling of both the content of the various models and relations between them, as per the section next.

All Supermodels implemented to date use Semantic Web modelling techniques and, specifically, the [RDF], [RDFS] and [OWL] fundamental models.

These models are given the role of Background Model within a Supermodel but are often not listed in overview diagrams, due to their low-level nature.

Model Relations

Models within the Supermodel relate to one another in various ways. The most common way is for one to profile another where profile means:

a specification that constrains, extends, combines, or provides guidance or explanation about the usage of other specifications [PROF]

Within Supermodel practice, profiling another model means using elements and patterns from it and specialising and extending them while ensuring that any data conforming to the profiling model still conforms to the profiled model.

For example, a model with the Integrative Model role might declare a class for Person and require that all Person instances have present one predicate of height indicating a float value. A profiling model would be free to declare a subclass of Person called Student with specialised predicates relating to student things, but it must still ensure that all instances of Student require a height predicate so that they are still valid Person instances.

Profiling between models within a Supermodel is indicated with the prof:isProfileOf predicate on the main model element, usually an instance of owl:Ontology.

Demonstrations

References

[PROF]

World Wide Web Consortium, The Profiles Vocabulary, W3C Working Group Note (18 December 2019). https://www.w3.org/TR/dx-prof/