Environmental Engineering Reference
In-Depth Information
The process used in the SEAMLESS project for creating a common ontology
for models, indicators and raw data was based on a participatory and collaborative
approach. A dedicated group of scientists was formed with participants from
all parts of the project. This cross-disciplinary group, named DOT.force (after
Data and Ontology Task force), aimed to develop a common ontology that
represents a shared conceptualization between the different databases, models
and indicators, involved in the SEAMLESS-IF process. The DOT.force listed
among its members, knowledge engineers and domain experts. Knowledge
engineers have the technical skills and relevant experience in ontology design
and conceptual modelling. The domain experts hold knowledge about a specific
domain, like a certain model or dataset, a set of indicators, a particular case
study or scenario, or even the SEAMLESS-IF process. Domain experts interacted
with knowledge engineers for specifying the conceptualizations involved in their
niches, and ultimately facilitated the development of specific parts of the ontology
(Janssen et al.
2009)
.
The knowledge engineers in the DOT.force worked on a number of actions that
lead to a complete and consistent ontology. These actions included:
1. Integrating the different database schemas into a single SEAMLESS database
schema;
2. Clarifying the model interface data structures (domain models), while adding
relevant metadata;
3. Associating indicators with model result sets or other static data collections;
4. Supplementing the ontology with meta-data on the concepts it holds, like
textual descriptions, data sources, other documentation references, units and
value ranges;
5. Developing an upper ontology to cover concepts relevant with the SEAMLESS-IF
process and execution.
Different methods were used to construct the ontology for the different actions.
For actions one and two on databases and models, dedicated meetings were
organized to develop the ontology, while for action three on indicators a proposal
was made by the knowledge engineers, which was then evaluated by relevant
domain experts. Action four on metadata was carried out independently by domain
experts, once agreement on the common ontology was reached between domain
members and knowledge engineers. For action five on project and scenario definition
an iterative process was used to develop a document, which was later synchronized
with the project ontology after each iteration.
The database design is directly mandated by the object structure of the entities
and properties described in the ontology and the data can be stored in the database
based on an adjunct persistence XML (eXtensible Markup Language) document
provided along with each class implementing the ontology structure.
Other working groups within the project have been populating the database
following the schema generated by the ontology that the DOT.force concluded to.
To explain more details about the knowledge management technology, we need first
to give an overview of the SEAMLESS-IF architecture.
