Information Technology Reference
In-Depth Information
transmission
radiator
engine
steering
Fig. 5.
Merging ontologies with communities
Definition 4.
Let
S
be a seed and
IK
(
S
) be a community generated by
S
. We call
dy-
namic seed generated by S
the set of positive answers given by a classifier trained
with S. We indicate this set with
Cl
(
IK
(
S
))).
Definition 5.
Let
S
be seed and let
Cl
(
IK
(
S
))) be the dynamic seed generated by S.
We call dynamic concept generated by S, and we indicate it with Dyn(S), the web
community dynamically generated by
Cl
(
IK
(
S
))). Formally:
Dyn
(
S
) =
Kl
(
Cl
(
IK
(
S
)))). (1)
Whenever there is an evolution in the connectivity of
IK
(
S
), two new seeds
S
' and
S
''
are generated by
S
. By classifying
IK
(
S
') and
IK
(
S
'') with
S
' and
S
'' as training sets,
we get two new concepts c' and c'', respectively.
Example 2
: Asbestos Removal meets Photovoltaic Roofs
We simply illustrate the procedure above by applying it to a specific case related to a
collective intelligence rooted in social networks focused on environmental issues.
The process is graphically described in Figure 6.
Let us assume that our collective intelligence contains a concept corresponding to
“Asbestos Removal” pointing to a specific community of Web sites. As is well-
known, many buildings contain asbestos, which was used in spray-applied flame-
retardant, thermal system insulation, and in a variety of other materials. Roofs of