Friday, April 29, 2005
On the discussion of data
regularity à
On the Stephenson Cyber Attack
Taxonomy à
Development of Judicial Process Ontology
and
Administrative Rulings over Harmonized Tariff Schedule (HTS) codes
The first application will serve to instance an ontology about the Harmonized Tariff Schedule administrative rules.
This ontology is not about the HTS itself; that is a interesting but secondary object of investigation. The ontology will be about the process of asking for and getting administrative rulings and the event space that is indicated by the publicly posted rulings:
The development of a Judicial Activity ontology is first, and will likely serve as one of several Upper DOF Abstract Ontologies.
DOF (Differential Ontology Framework) has three
layers of ontology.
1: the upper abstract
ontologies
1.1:
no concept subsumption is allowed because
1.1.1: the abstract nature of the concepts in the upper DOF ontology
1.1.2: a technical reason related to a need to propagate changes made
to the upper DOF within the distributed Ontology Management Architecture
(d-OMA).
1.1.3: an ecstatic reason related to the way in which middle layer
ontology refers to concepts that are specified in the upper DOF.
1.2:
only concepts needed by more than one middle ontology are placed into the upper
DOF, and only then when there is a framework-based reason for this inclusion.
1.2.1: a framework is an organizational “device” that uses categories
as a basis for not recognizing the differences in structure (syntax) when there
are functional equivalences. Example: (see PowerPoint: slide 2) The counting numbers have elements that do
not care if we have “two” apples or “two” oranges.
1.2.2: Stephenson and Prueitt’s
FARES ontology has an upper DOF ontology about (communication) Risks in
distributed government enterprises.
1.2.3: Deductive inferencing is minimized due to the generality of the
concepts and because there are no hierarchical relationship in upper DOF
ontology.
2: Middle DOF ontology
2.1:
Middle ontology is domain and system specific, but does not focus on
events.
2.2:
Middle ontology is generally, but not always, carefully guarded by the
enterprise because domain ontology gives the specifics of how an organization
is known to work.
2.3:
Middle ontology concepts can have subconcepts and deductive (RDF) inferencing
3: Scoped Ontology Individuals
3.1:
By “individual” we mean that this ontology is about a specific event.
3.2:
The Scoped Ontology Individual is developed by allowing a semantic extract
process to act on a event report, such as one of the HTS administrative
rulings. The result of the semantic
extract process is an ontology expressed as an Orb construction, i.e., having
the form of a set of triples, { < a, r, b > }. Following the computational semiotics literature, this triple is
called a syntagmatic triples, and is deemed to not be about meaning – but
rather about structure.
3.2.1: This Orb construction is streamed to various destinations depending
on a contact list and a set of rules.
This allows real time monitoring over characteristics that might need to
be monitored, i.e. oranges if there is an particular alert for a bioterrorism
attack.
3.2.2: The Orb is used to
produce an expansion of information available to the SOI. This new information is added to the Orb in
the form of new syntagmatic triples.
3.2.3: The Orb is used to produce an contraction into a Framework. The contraction does have conjectures about
the meaning of things. Example, the
event Structure Ontology Framework attempts to compress all information about
an event into answered to 18 questions.
(see PowerPoint; Slides 8 and
9)
3.2.4: Visualization of the SOI
is made directly to the user, who can produce modifications to the graph,
labels or properties of the Orb. After
the user has completed review, use, and modification; the SOI is moved to a repository
where global event analysis is performed.
3.2.5: RDF repositories (middle DOF and upper DOF ontology) can be
modified based on the analysis of SOI.