=Paper=
{{Paper
|id=Vol-210/paper-12
|storemode=property
|title=Ontology Verification Using Contexts
|pdfUrl=https://ceur-ws.org/Vol-210/paper12.pdf
|volume=Vol-210
|dblpUrl=https://dblp.org/rec/conf/ecai/SegevG06
}}
==Ontology Verification Using Contexts==
https://ceur-ws.org/Vol-210/paper12.pdf
Ontology Verification Using Contexts
Aviv Segev and Avigdor Gal
Technion – Israel Institute of Technology
{asegev@tx, avigal@ie}.technion.ac.il
Abstract. Ontologies have become the de-facto modeling tool of to be a set of words, possibly associated with weights that represent
choice, used in a variety of applications and prominently in the Se- the relevance of a word to a document. Ontologies and contexts are
mantic Web. Their design and maintenance, nevertheless, have been both used to model different perspectives of a domain (views). On-
and still are a daunting task. As a result, ontologies quickly become tologies represent shared models of a domain and contexts are local
underspecified. Therefore, if ontologies do not evolve, the semantic views of a domain. We also promote an orthogonal classification in
infrastructure of the information system can no longer support the which ontologies are considered a result of a manual effort of mod-
changing needs of the organization. In this work we provide a model eling a domain, while contexts are system generated models [8]. On-
to semi-automatically support relationship evolution in an ontology tologies and contexts are joined together, as formally described in
using contexts. We propose to use (machine-generated) contexts as a [9]. In a nutshell, each concept in an ontology is represented by a
mechanism for quantifying relationships among concepts. To do so name and a context. In this model, contexts serve as an easy-to-use
we compare the contexts that are associated with the ontology con- “semantic glue,” in which underspecifications are compensated for
structs. On a conceptual level, we introduce an ontology verification with a syntactic, machine generated context, which highlights the in-
model, a quantified model for automatically assessing the validity of tentions of a local designer when using a specific ontology concept,
relationships in an ontology. We motivate our work with examples possibly differently from the way it is semantically captured in the
from the field of eGovernment applications. To support our model ontology using relationships.
with an empirical analysis, we provide a mapping of an ontology In this work we provide a model and an example of an algorithm
operator for defining relationships into context relationships, using to semi-automatically support relationship evolution in an ontology
real-world traces of RSS. using contexts. The main motivation for this work stems from the
difficulty in supporting ontology evolution. Specifically, this problem
was raised within the framework of TerreGov, a European eGovern-
1 INTRODUCTION ment project. In this project, ontologies serve as the driving force
behind the application and thus affect government processes and
Ontologies have become the de-facto modeling tool of choice, used
Web services, among other things. Therefore, we propose to use
in a variety of applications and prominently in Semantic Web ap-
(machine-generated) contexts as a mechanism for quantifying rela-
plications. For example, ontologies can be used in discovering Web
tionships among concepts. Specifically, given an ontology operator
services [10]. Ontology design, nevertheless, has been and still is a
(e.g., link subclass, representing the knowledge that an instance of
daunting task. It requires collaboration of domain experts with ontol-
one concept is included in an instance of another) and operands (e.g.,
ogy engineers, which may consume many organizational resources
two concepts or classes), we aim at quantifying the extent to which
in terms of both time and monetary units. Once the ontology is de-
this relationship is valid. We do so by comparing the contexts that are
signed, evolving it becomes difficult due to the need for availability
associated with the operands. We believe that such a solution would
of domain experts on the one hand, and costs related with hiring on-
significantly assist in the support of ontology design and evolution.
tology engineers on the other hand. To illustrate this point, consider
The main contribution of this work is thus twofold. On a concep-
an eGovernment application, for which an ontology was designed
tual level, we introduce an ontology verification model, a quantified
and tailored by an ontology engineer. Once the ontology is installed,
model for automatically assessing the validity of relationships in an
changes in the real world require a renewed collaboration of civil
ontology. On an algorithmic level, we provide an example of a map-
servants with ontology engineers to reflect such changes in the ontol-
ping of ontology operator for defining relationships into context re-
ogy. A typical outcome of such difficulties is that ontologies quickly
lationships. We motivate our work with examples from the eGovern-
become underspecified. New concepts are introduced in the domain
ment domain. However, due to the absence of large scale data sets for
while others become obsolete. Also, shifts of focus in the application
this domain, we support our model with an empirical analysis using
domain require the refinement of a concept into a hierarchy of con-
real-world traces of RSS data.
cepts, while in other cases hierarchies should be collapsed. Meeting
The rest of the paper is organized as follows. We start with pre-
these challenges requires ontologies to evolve or else the semantic
liminaries, formally defining ontologies and contexts in Section 2.
infrastructure of the information system can no longer support the
In Section 3 we introduce the ontology verification model, followed
changing needs of the organization.
by an example of a mapping of the ontology verification problem to
In [9] we introduced a model for compensating for ontology un-
contexts in Section 4. We conclude with related work in Section 5
derspecification using a combination of ontologies with contexts.
and a short summary in Section 6.
Contexts were defined to be first class objects [5] and will be formally
presented later in this work. As an example, a context can be defined
2 ONTOLOGIES AND CONTEXTS
Banerjee [1] defined a root class as an object that represents any-
thing from a simple number to a complex entity. An edge between
a node and a child node in a class represents an IS-A relationship.
Objects that belong to a class are called instances of that class. A
Ontology
class describes the form (instance variables) of its instances and the Q
operations (methods) applicable to its instances.
According to Gruber [2], an ontology is an explicit specification of Verification
a domain conceptualization. Several models for ontologies exist; we Operator Operand Certainty Level
follow here that presented in [2]. In the discussion below, we assume Disjoint CL1, CL2
0.9
reader familiarity with basic concepts
in conceptual modeling.
We define a context C = {hcij , wij i}j i as a set of finite sets
of descriptors cij from a domain D with appropriate weights wij ,
Figure 1. Ontology Verification Model
representing the importance of cij . For example, a context C may be
a set of words (hence, D is a set of all possible character combina-
tions) defining a document Doc, and the weights could represent the
relevance of a descriptor to Doc. In classic Information Retrieval,
4 EXPERIENCES WITH CONTEXT BASED
hcij , wij i may represent the fact that the word cij is repeated wij
ONTOLOGY VERIFICATION
times in Doc.
The context of a class is defined as a set of contexts describing
Having introduced ontology verification, we now focus on the details
instances that belong to this class. Documents are not instances but
of change operators. Noy and Klein [6] describe a set of 22 ontology
represent them. Following [9], we define a class context CCL of a
change operators and their impact on ontology elements (both classes
class CL to be the union of its instance contexts.
and instances) using ontology relations defined in [2]. We take one
Segev and Gal [9] aimed at formalizing the inter-relationships be-
of their ontology change operators and use it as an example.
tween an ontology, a manually generated domain model, and con-
Our experiences are based on data from the RSS news data trace.
texts, partial and automatically generated local views. According to
In this data trace, data were originally partitioned to topics with no
their work, a context can belong to multiple context sets, which in
ontological relationships. The RSS trace was collected during August
turn can converge to different ontology concepts. Thus, one context
2005 from the CNN Web site. We chose 10 news topic categories for
can belong to several ontology concepts simultaneously. The appro-
the data, when each RSS news header or news descriptor constitutes
priate interpretation of a context leads to its relevance to different
a datum. We generated a context for each datum and each class using
given concepts.
an automatic context extraction algorithm [8]. The number of context
descriptors generated from each datum was set to 10. The data size
used for RSS varied from 73 to 1,911 per class.
3 ONTOLOGY VERIFICATION USING In our experiment we calculated for each class the number of con-
CONTEXTS texts that overlapped with the other nine classes. This asymmetric
comparison gave us for any two classes CLi and CLj the metric of
Ontology verification is the process by which semantic relationships CCLi ∩ CCLj and CCLi ∪ CCLj .
are identified. We term this process verification, since we assume Given two classes, CLi and CLj , if CLi is a subclass of CLj ,
an ontology exists and may need to evolve. Therefore, semantic re- then its context should be contained in the context of CLj . This is
lationships in an ontology need to be continuously monitored and because an instance of CLi is also an instance of CLj and therefore
if necessary, revised. Here we follow the work of [6] on ontology has a broader context than an instance of the superclass. Therefore,
changes and assume a given closed set of operators OT , to be ap- we compute the certainty of a hypothesis that CLi is a subclass of
plied on a set of operands OD, taken from the set of all ontology CLj to be
elements. As an example, a change operator may be the disjoint op- CCLi ∩ CCLj
erator, resulting in the creation of a semantic relationship called “dis- µSub-Sup =
CCLj
joint” between two classes, given to it as operands.
Figure 1 provides a pictorial representation of the process. For- Our experience involves an analysis of hierarchy linking. Figure 2
mally, ontology verification is a function OV : OT ×OD∗ → [0, 1]. presents the RSS class relations hierarchy created for µSub-Sup ≥
Ontology verification is given as input a hypothesis regarding the 0.8 and µSub-Sup ≥ 0.5. As the value of µSub-Sup decreases, the
possible operator to be applied to one or more operands and returns
hierarchy and the relations between the classes become more elab-
a level of certainty µ regarding the truth in this hypothesis. A cer-
orated. For example, in the RSS data for µSub-Sup ≥ 0.8 the su-
tainty of 1 indicates full certainty in the hypothesis, while a certainty
of 0 means that the hypothesis is definitely incorrect. In Figure 1, the perclass Money Latest has four subclasses. If we examine the same
ontology verification function determines that the disjointedness of classes for a lower verification level of µSub-Sup ≥ 0.5 we receive
classes CL1 and CL2 has a certainty level of 0.9. An example of a three level hierarchy.
the use of the model can be a user who would like to analyze a local Table 1 compares the certainty level of the Superclass-Subclasss
government concept relationship. The user could supply a set of doc- operator, for two class pairs in the RSS data set. When evaluating the
uments representing two concepts and could receive a verification classes Money Latest and Money News International, there is a
level based on this representative set of documents. high µSub-Sup level.
Class Sets Link Subclass
Money Latest 86.7%
Money News International 19.8%
Money News Economy 19.5%
Money Latest
sub-sup 0. Money Markets 24.3%
Table 1. Operator µ Verification RSS
Money News Money Money News Money Top
Economy Markets International Stories
Given an ontology operator and operands, the model provides the
quantification of the extent to which the relationship is valid. The
model is supported by empirical analysis, using initial experiences
with real-world RSS traces. The experiences with these traces show
Money Latest
sub-sup 0. how relationships between the classes can be created and modified.
Preliminary empirical results show that our model can provide good
estimations of the need for ontology changes.
To recap, the main contribution of this work is both conceptual
Money News Money Top and algorithmic. We present an ontology verification model, a quan-
International Stories
tified model for automatically assessing the validity of relationships
in an ontology, and we also provide a mapping of several ontology
operators for determining relationships among classes.
The results of this work will be embedded as part of the Terre-
Money News Money Gov solution. Future research will examine the model performance
Economy Markets
on eGovernment data and other large data sets. In addition, we plan
on extending the model to include additional operators.
REFERENCES
Figure 2. RSS Relations [1] J. Banerjee, H.-T. Chou, J. Garza, W. Kim, D. Woelk, and N. Ballou,
‘Data model issues for object-oriented applications’, ACM Transactions
on Office Information Systems, 5(1), 3–26, (1987).
[2] T. R. Gruber, ‘A translation approach to portable ontologies’, Knowl-
edge Acquisition, 5(2), (1993).
5 RELATED WORK [3] K. Knight and S. K. Luk, ‘Building a large-scale knowledge base for
A formal mathematical framework that delineates the relationships machine translation’, in Proceedings of the Twelfth National Confer-
ence on Artificial Intelligence (AAAI-94), (1994).
between contexts and ontologies is presented in [9]. To deal with the [4] D. B. Lenat, ‘Cyc: A large-scale investment in knowledge infrastruc-
uncertainty associated with automatic context extraction from exist- ture’, Communications of ACM, 38(11), 33–38, (1995).
ing instances, such as documents, a ranking model was provided, [5] J. McCarthy, ‘Notes on formalizing context’, In Proceedings of the
which ranks ontology concepts according to their suitability with a Thirteenth International Joint Conference on Artificial Intelligence,
(1993).
given context. [6] N. F. Noy and M. Klein, ‘Ontology evolution: Not the same as
A semi-automated method for ontology evolution using docu- schema evolution’, Knowledge and Information Systems, 6(4), 428–
ments clustering was proposed in [11]. From the results of the clus- 440, (2004).
tering ontology enrichments and updates are extracted. In contrast to [7] N. F. Noy and M. A. Musen, ‘The prompt suite: Interactive tools
the above work, which is based on a single word ontology concept for ontology merging and mapping’, International Journal of Human-
Computer Studies, 59(6), 983–1024, (2003).
description, we use a set of contexts describing each ontology class. [8] A. Segev, ‘Identifying the multiple contexts of a situation’, in Pro-
Noy and Klein [6] defined a set of ontology-change operations ceedings of IJCAI-Workshop Modeling and Retrieval of Context
and their effects on instance data used during the ontology evolution (MRC2005), (2005).
process. They describe ontologies schemas and database schemas [9] A. Segev and A. Gal, ‘Putting things in context: A topological approach
to mapping contexts and ontologies’, in Proceedings of AAAI-Workshop
from the point of view of evolution and highlight the main differ- Workshop on Contexts and Ontologies: Theory, Practice and Applica-
ences between them. We presented a model that shows how these tions, (2005).
ontology change operations can be verified based on context. [10] E. Toch, A. Gal, and D. Dori, ‘Automatically grounding semantically-
Tools for merging and aligning ontologies, such as SENSUS [3], enriched conceptual models to concrete web services’, in ER, eds.,
PROMPT [7], and Cyc [4], have been developed in the past. These L.M.L. Delcambre, C. Kop, H.C. Mayr, J. Mylopoulos, and O. Pas-
tor, volume 3716 of Lecture Notes in Computer Science, pp. 304–319.
tools generally present a set of basic operations that are performed Springer, (2005).
during the mergence and alignment of ontologies and that determine [11] G. Tsatsaronis, R. Pitkanen, and M. Vazirgiannis, ‘Clustering for on-
the effects that each of these operations has on the process. tology evolution’, in Proceedings of the 29th Annual Conference of the
A work on multi-contextual ontology evolution [12] defines a set German Classification Society (GfKl 2005), (2005).
[12] M. Zurawski, ‘Reasoning about multi-contextual ontology evolution’,
of properties that by semantic autonomy must hold at the same time. in Proceedings of the First International Workshop on Context and On-
tologies: Theories, Practice and Applications, The Twentieth National
Conference on Artificial Intelligence (AAAI-05), (2005).
6 CONCLUSION
This work presents a model and a set of algorithms to semi-
automatically support ontology relationship evolution using contexts.