Information Technology Reference
In-Depth Information
Waffles were ordered by John—was-ordered-by
(the Belgian Waffles, John)” (3). This leads to
the tendency in RDF to create redundant, inverse
relationships (of which (1) and (3) are examples).
Yet, the ability (provided by DAML-OIL) to state
explicitly that “ordered” and “was-ordered-by” are
inverse relationships does not completely solve
the redundancy problem. In Topic Maps, it is not
possible to assert that “John ordered the Belgian
Waffles” without also asserting that “The Belgian
Waffles were ordered by John”; they are one and
the same association (Pepper, 2002b).
This additional expressivity is made possible
by the notion of association roles, which make
clear the kind of role played by each participant in
a relationship. The association role can illustrate
more than binary relationships (Pepper, 2002b).
In RDF, assertions are always binary. An RDF
statement, consisting of a subject, a predicate,
and an object, expresses a relationship between
subject and object, for example, “John ordered
the Belgian Waffles—ordered (John, the Belgian
Waffles).” In Topic Maps, assertions are n-ary
(Pepper, 2002b). An association may have any
number of roles and can thus represent more com-
plex relationships (Pepper, 2002b), for example,
“John ordered the Belgian Waffles with maple
syrup—ordered (John, the Belgian Waffles, maple
syrup).” Understanding the differences between
Topic Maps and RDF allows information profes-
sionals to choose either Topic Maps or RDF for
their metadata migration. Specifically, if data has
more than binary relationships, Topic Maps will
be a better choice to migrate data.
dependent variables, recall and search time, were
measured. The findings of this study show that a
Topic Maps-based ontology information retrieval
(TMIR) system has a significant effect on both
recall and search time, compared to a thesaurus-
based information retrieval (TIR) system.
The experiment purposely examined associa-
tive relationships between resources belonging to
different hierarchies that are explicitly provided
and could be recognized as better candidates
for improved recall and shorter search time.
This study demonstrates that relationship-based
query searches using this TMIR system resulted
in improved recall and shorter search times than
fact-based query searches. The results of this study
demonstrate the possibility of Topic Maps-based
ontology to enhance information retrieval system
performance through better support for associa-
tive relationships between resources belonging
to different hierarchies by providing explicit
relationships among resources.
Significant difference in recall and search
time was found between the experimental group
performing fact-based queries and the control
group performing fact-based queries. As illus-
trated in Figure 4, the average recall percentage
for the experimental group performing fact-based
queries was 83%, and the average recall percent-
age in the control group performing fact-based
queries was 85%. The average search time for the
experimental group performing fact-based queries
was 91 seconds, and the average search time for
the control group performing fact-based queries
was 86 seconds (See Figure 5).
One of the most significant findings of this
study was the substantial difference in search time
and recall between the performance of relation-
ship-based queries using a TMIR and the perfor-
mance of relationship-based queries using a TIR
system. There was a significant mean difference
in recall and search time between the experimental
group performing relationship-based queries and
the control group performing relationship-based
queries. As illustrated in Figure 6, the average
user PerFormAnCe using
toPiC mAPs-bAsed ontology
inFormAtion retrievAl system
The recent study by Yi (2008b) shows user
performance using a Topic Maps-based ontol-
ogy information retrieval system. Forty subjects
participated in a task-based evaluation where two
Search WWH ::
Custom Search