Back ...
... ... ... ... ... ... ... ... ... ... Send comments into the ontology
stream. ... ... ... ... ... ... ... ... ... ... ... Forward
(:..:) ß different threads à (:….:)
Bead 2
January
1st, 2001
OntologyStream Fable Indexing Project
Bead 2
January 6th, 2001,
Each passage within
a fable might have an image. These
images might be placed on playing cards.
The cards might then be sorted into as many piles as someone might
wish. We need an artist to represent
all of these images.
This procedure will
identify passage categories. It is
similar to the method of descriptive enumeration. This method is presented in slides 6 – 8 in a presentation made
to e-Gov conference on Sense
Making in 2000.
The method is the
very core to knowledge science, in our opinion.
Each pile of cards is
selected to be a “passage category”.
The group of categories (called a category policy in the tri-level
voting procedure) represents the semantic representations that are the invariance
of the collection.
Notice that the
category representation is dependant on the selection process conduced by human
hands and visual acuity. The ground
truth for the OntologyStream
IRR Test is developed by human judgment.
The question is now
about the relationship between the categories.
These categories can be regarded as locations in a space and some sort
of latent semantic indexing used to acquire meaning.
However, the ground
truth for the relationships between categories can also be vetted by human
judgment. In this way, the use of
algorithms can be measured against this human judgment.
If two fables each
have a passage category in common, then there is a specific semantic linkage
that now is said to exist. In every
case where a passage category is shared between two passages, a literature
scholar will provide a descriptive phrase for the relationship.
Thus we develop a
collection of syntagmatic units { < a, r, b > } where a and b are
locations in a semantic space and r is a relationship variable.