KYOTO Logo
Knowledge Yielding Ontologies
for Transition-based Organization
  • Increase font size
  • Default font size
  • Decrease font size
KYOTO Project Forum
Welcome, Guest
Please Login or Register.    Lost Password?
Non-rigid terms, types and roles (1 viewing) (1) Guests
Go to bottom Post Reply Favoured: 0
TOPIC: Non-rigid terms, types and roles
#41
AmandaHicks (User)
Fresh Boarder
Posts: 9
graph
User Offline Click here to see the profile of this user
Non-rigid terms, types and roles 2 Years ago Karma: 3&nbsp&nbsp
There are two sections to this post.
Comments and suggestions are welcome!


I. Modelling roles in the KYOTO ontology
II. Adding a role to the ontology _base_d on the domain wordnet

I. Modelling roles in the KYOTO ontology

The role hierarchy will be entered into the ontology, under the node “concept”. The relation between endurants and roles is the “played-by” relation, which is a subclass of the “classifies” relation. Hence, the role “pet” can be played by “animal” or an instance of animal, i.e., Fluffy. The inverse relation of “played-by” is “plays”, i.e., endurants play roles.

Fluffy is an instanceOf plays some pet

Since roles are always modelled as concepts in this ontology, there are two cases to consider when modelling relations with roles. A concept-to-concept relation, and a concept-to-individual role relation.

The following examples, taken from DOCLE-Lite-Plus, illustrate a concept-to-concept relation.

• substance-role subClassOf (played-by only amount-of-matter)
• de_script_ion-role equivalentTo role and played-by only de_script_ion

In the first example, “substance-role” is given the parameter that it is a subclass of those roles that can only be played by an amount of matter. In the second example, “de_script_ion-role” is defined as a role that is only played by de_script_ions.

The relations between roles and individuals as represented by defining concepts using the “plays” relation and stating that the individual is an instance of this concept. Here are two examples.

• Fluffy instance-of (plays some pet)
• Tree frog instance-of (plays some invasive_species)

The first example states that Fluffy is an instance of those things that play that role “pet”. Likewise, the second example states that tree-frog is an instance of those things that play the role of an invasive species.

Notice that if further kinds of invasive species are later added to the role hierarchy, then this axiom is consistent with these extensions.

Finally, we should note that, although certain roles in DLP are treated as individuals, KYOTO will steer clear of this approach. The following example is taken from DLP.

• meaning instanceOf (played-by only de_script_ion)

This states that the individual ‘meaning’ is an instance of those roles that are only played by de_script_ions. However, given the fact that the KYOTO ontology is designed to be continually expanded, we will avoid this approach. Once a role is entered as an individual, there cannot be further sub-roles of this role. For example, if “invasive species” appears in the ontology as an individual, then it is not clear how to model the concept “tropical invasive species”. It certainly cannot be given a hierarchical relation with “invasive species” if “invasive species” is an individual. Therefore, in order to leave open the possibility of further subroles, we choose to model roles as concepts by default.


II. Adding a role to the ontology _base_d on the domain wordnet.

Moving to the role hierarchy, decision tree

Is it a sortal? (We make the general assumption that nouns express sortals.)
No. Then it is neither a type nor role.
Yes. Continue
Is it an endurant?
No. Then it does not belong in the role hierarchy.
Yes. Is it a de_script_ion or a situation?
Yes. Then it does not belong in the role hierarchy. Leave it.
No. Move it to the role hierarchy (make a set of instructions for this).

Moving a wordnet node to the role hierarchy

The role hierarchy is structured according to the domain of the played-by relation for each role represented in the hierarchy. For example, a species role is played by a species; a taxonomic group role is played by a taxonomic group, and since taxonomic-group subsume species in the role hierarchy, taxonomic-group-role subsumes species-role in the role hierarchy. This structure in the role hierarchy is largely heuristic to help the user navigate and locate a role.

When it has been determined that a term in a wordnet represents role, then it can be moved to the role hierarchy.

Suppose this were a hierarchy in the domain wordnet.

Species - Type
Invasive species - role
Tropical invasive species – role

In order to move “invasive-species” to the role hierarchy, find the hypernym, in this case “species”, and then locate the corresponding role in the role hierarchy, in this case “species-role”. (If it does not already exist, it can be automatically created.) Then create an instance relation between “invasive species” and “role”.

When creating the species role, we need to create a restriction that specifies that species roles are played by species.

• species-role equivalentTo role and played-by only species

Because ‘tropical-invasive-species’ is subsumed under ‘invasive-species’ in the wordnet, it should also be modelled as a role in the ontology. Roles can only subsume roles. They cannot subsume types.
 
Report to moderator   Logged Logged  
  The administrator has disabled public write access.
      Topics Author Date
    thread link
Non-rigid terms, types and roles
AmandaHicks 2009/04/22 18:34
Go to top Post Reply
Powered by FireBoardget the latest posts directly to your desktop

ICT-211423 - 2008 © Kyoto Consortium