=Paper=
{{Paper
|id=None
|storemode=property
|title=FCA-Based Concept Detection in a RosettaNet PIP Ontology
|pdfUrl=https://ceur-ws.org/Vol-1058/paper3.pdf
|volume=Vol-1058
|dblpUrl=https://dblp.org/rec/conf/ijcai/JridiL13
}}
==FCA-Based Concept Detection in a RosettaNet PIP Ontology==
https://ceur-ws.org/Vol-1058/paper3.pdf
FCA-Based Concept Detection in
a RosettaNet PIP Ontology
Jamel Eddine Jridi and Guy Lapalme
DIRO, Université de Montréal, Canada,
{jridijam, lapalme}@iro.umontreal.ca
Abstract. This paper presents an FCA-based methodology for concept
detection in a flat ontology. We apply this approach to an automatically
generated ontology for a RosettaNet Partner Interface Process (PIP)
which does not take advantage of some important OWL semantic rela-
tions like subClassOf. The goal of our approach is to regroup ontology
classes sharing a set of properties (Data and Object Property) in order
to improve the quality, readability and the inheritance richness of a flat
ontology.
Keywords: Formal Concept Analysis, Ontology, RosettaNet PIP On-
tology, Inheritance Richness.
1 Introduction
Ontologies are widely used in knowledge management, information integration,
natural language processing, information retrieval, business-to-business (B2B),
e-commerce, etc [4].
Research is now envisioning the adoption of semantic web technologies in the
business domain. Lytras et al. [8] analyse the practical requirements in terms of
interoperability or knowledge representation. The use of ontologies is not only
for communication between different applications but also to provide reasoning
support to infer, integrate information and to extract meaning.
After an ontology is constructed, it is important to assess the quality of
an ontology to detect design defects and to automatically recognize parts that
cause problems and might need more work. In this paper, we try to improve the
inheritance richness of an OWL ontology based on the balance between depth
and height of the inheritance tree which, according to Tatir et al. [13], play a
role in a quality assessment. In Software Engineering, Sheldon et al. [10] claim
that when the hierarchy in the inheritance tree is shallow, better will be the
maintainability, readability and understanding.
In this paper, we propose a method to verify, regroup concepts sharing prop-
erties to improve the quality and readability of an automatically generated on-
tology. We use Formal Concept Analysis in the context of concept detection in
OWL ontologies using OWL artifacts (Class, DataProperty and ObjectProperty).
The remainder of this paper is structured as follows. Section 2 states the
problem. In Section 3, we describe the principles of the Formal Concept Anal-
ysis algorithm underlying our approach and the adaptation of its principles for
the detection of concepts in a RosettaNet Ontology. Section 4 presents and dis-
cusses the validation results. A summary of the related work in ontology-based
attempts of B2B standards and the use of Formal Concept Analysis in Ontolog-
ical Engineering is given in Section 5. We conclude and suggest future research
directions in Section 6.
2 Background and Problem Statement
In this section, we describe the problem of concepts detection and the importance
of the inheritance tree to represent a domain knowledge. We start by defining
important notions. Then, we detail the specific problems that are addressed by
our approach.
2.1 Basic notions
Ontology in Semantic Web technology provides a shared and common vocab-
ulary for a domain of interest and a specification of the meaning of its terms [3].
It allows users to organize information into a hierarchy of concepts, to describe
relationships between them, and to make semantics machine processable, not
just readable by a human.
RosettaNet B2B Standard is a consortium which provides a global forum
for suppliers, customers, and competitors to do business and collaboration in
an efficient and profitable manner. To manage business activities, RosettaNet
formalizes Partner Interface Processes (PIP) with either Data Type Definition
(DTD) format or XML Schema, and define business processes between trading
partners. PIPs are organized into eight groups of core business processes called
clusters, themselves further grouped into segments. Each segment includes sev-
eral PIPs. In section 3, we will use the PIP3A4 as running example of our
methodology. But we managed to apply the same technique on all published
PIPs.
The RosettaNet architecture contains a Business Dictionary to define the
properties for basic business activities and Technical Dictionaries to provide
properties for products [14]. The RosettaNet Implementation Framework (RNIF)
describes the packaging, routing, and transport of all PIP messages and business
signals.
2.2 Problem Statement
There are few works to measure quality of an ontology using metrics. But, to
our knowledge, no work has yet proposed methods to verify, maintain concept
hierarchy representation and improve the quality and readability of an ontology
from inheritance point of view using OWL artifacts (Class, DataProperty and
ObjectProperty).
Tatir et al. [13] and Sicilia et al. [11] propose an Inheritance Richness metric
defined as the average number of subclasses per class and represent the dis-
tribution of information across different levels of the ontology inheritance tree.
Values close to zero indicate horizontal ontologies representing perhaps more
general knowledge while large values represent vertical ontologies describing de-
tailed knowledge of a domain.
In our case study, we have chosen an ontology describing a detailed domain
knowledge. Some approaches are proposed to ontologize some of the famous B2B
standards like RosettaNet1 and ebXML2 .
In this paper, we apply our methodology on a RosettaNet Ontology described
in our paper published in the 15th International Conference on Enterprise Infor-
mation Systems [5]. RosettaNet focuses on a supply chain domain and defines
processes and business documents in detail. This automatically generated on-
tology, described in Section 4, has some drawbacks because it does not take
advantage of some important OWL semantic relations like subClassOf. It is
quite flat and does not describe in details the supply chain knowledge provided
by RosettaNet, although it deals with a specific domain having several concepts
with a common semantics. Although approaches have been previously proposed
to ontologize a few B2B standards like RosettaNet and ebXML [2,6,7], we ar-
gue in [5] that our methodology is the first to deal with the complete set of
RosettaNet PIPs.
During the transformation process, it is difficult to automatically detect el-
ements with a common semantics which are described in separate OWL files.
But, we noticed that many of those share some properties and lexemes in their
compounded names. For this reason, we use FCA as a classification approach to
detect concepts by regrouping classes in an ontology to improve its quality and
readability.
3 FCA-Based Concept Detection Approach
The goal of our approach is to regroup ontology classes sharing a set of properties
(Data and Object Property) and then maintain the hierarchy representations.
The core of our system has three main parts: Ontology Processing, FCA Sys-
tem and Regrouping concepts. The Ontology Processing step builds a cross-table
from ontology artifacts (Class, Data Property and Object Property) without
dealing with the property type or occurrence restrictions. This table describes
relationships between objects (ontology classes are represented by rows) and
attributes (Data and Object properties correspond to columns).
Taking the example in Table 1, o1..k represents ontology classes with k is the
number of classes in the ontology. p1..m represents the set of Data and Object
1
http://www.rosettanet.org/
2
http://www.ebxml.org/
Properties with m being the total number of data and object properties. Element
of the table is marked with “x” if the domain of property pj is class oi .
We use FCA to create a hierarchy of concepts displayed as an inheritance tree.
According to [10], maintainability, readability and understanding of a hierarchy
are better when it is shallow. For this reason, we consider only two levels in the
hierarchy of the concept lattice.
Table 1. Example of cross-table.
R p1 p2 p3 .... pm
o1 x x x x
o2 x x x
...
ok x x x
6+13=19
6+7=13
7
Fig. 1. An example of a concept lattice describing RosettaNet PIP3A4 Purchase
Order Request. The process to build this concept lattice starts by extracting on-
tology artifacts (Class, Data Property and Object Property) from the OWL File of
PIP3A4 which are added to a cross-table as CSV file. This file serves as input to the
FCA system to build this concept lattice.
The concept lattice in Figure 1, which is the output of our FCA applica-
tion, describes the RosettaNet PIP3A4 (Purchase Order Request). Intermediate
concepts (of the form Concept_NN) were generated by the algorithm and rep-
resent the shared set of properties. The others represent the original ontology
classes.
Each class will be associated with the concept having the largest num-
ber of shared properties. As shown in Figure 1, we note that the Concept_7
combines the two classes ProductSubLineItemType and ProductLineItemType
with 19 shared properties; Concept_8 joins ServiceLineItemByOptionType and
ServiceLineItemType because of 22 shared properties.
The combined classes share some semantics consistent with the concept def-
initions provided by RosettaNet. They represent one concept (Concept_7 and
Concept_8) that has been generated by our approach. As the regrouped classes
share tokens in their compound names, we will use these to rename the FCA
generated names. So, Concept_7 becomes ProductLineItem and Concept_8,
ServiceLineItem.
Using the generated group of classes, we update the original ontology. To do
this, we used the OWL API3 for manipulating, developing and maintaining the
ontology.
4 Experimentation
The goal of our study is to evaluate the efficiency of our approach for concept
detection in an OWL ontology. In this section, we describe our experimental
setup and results.
In order to test our methodology, we use the RosettaNet Ontology, proposed
in our ICEIS 2013 paper [5]. It is the result of mapping the full set of Roset-
taNet Partner Interface Process (PIP) descriptions, currently defined with DTD
or XML Schemas format, to an ontological representation using an OWL/XML
rendering. Among the 132 PIPs, 112 PIPs are available for download from the
PIP Directory in RosettaNet website from which we generated 138 OWL doc-
uments, valid according to the XML Schema for OWL/XML serialization from
syntactical point of view. These OWL documents were also checked for consis-
tency with an OWL reasonner.
In Table 3, we evaluate our RosettaNet Ontology using state of the art ontol-
ogy metrics defined in Table 2. Only basic metrics related to the main elements
of ontologies have been used [11]. Although, we use two metrics that indicate
the relationship and inheritance richness of an ontology schema. The Relation-
ship Richness metric (rr ) reflects the diversity and placement of relations in
the ontology [13]. It is defined as the ratio of the number of properties (nop)
divided by the sum of the number of subclasses (nosc) plus the number of prop-
erties. Also, the Inheritance Richness metric (ir ) represents how knowledge
is grouped across different levels of the ontology inheritance tree [13]. It is the
average number of subclasses per class.
3
http://owlapi.sourceforge.net/
Table 2. Ontology metrics.
Metrics Definition
classes (noc) number of classes (|C|) in the ontology.
data properties (nodp) number of data properties.
Number of
object properties (noop) number of object properties.
properties (nop) sum of nodp and noop metrics.
subclasses (nosc) number of subclasses (|sC|) in the ontology.
number of root classes (without superclasses). The range
root classes (norc)
of this metric is between 1 and |C|.
classes without subclasses. The range of this metric is
leaf classes (nolc)
between 1 and |C|.
Table 3. Empirical analysis of our RosettaNet Ontology using some Ontology Metrics:
Before and After applying our FCA-based Concept Detection Approach.
Metrics All PIPs (Before) All PIPs (After)
noc 1252 1252
nodp 3045 3045
noop 2607 2607
nop 5652 5652
nosc 0 384
norc 1252 1050
nolc 1252 868
rr 1 0.93
ir 0 0.31
We notice in Table 3 that the values of ir and nosc metrics are zero and
the values of norc and nolc are equal to the number of classes |C| in the on-
tology. These metric values indicate that our original RosettaNet PIP Ontology
as generated from the DTD and XML Schemas is a flat or horizontal ontology
representing a general knowledge despite the fact that RosettaNet represents a
specific domain of interest with many elements sharing a common semantics.
The application of our FCA-based Concept Detection approach extracts 182
groups of concepts comprising 384 classes from the 1252 in the ontology, bringing
the inheritance metric ir from 0 to 0.31. On average, each concept contains 2
classes.
We extracted 90 groups of concepts because several concepts are shared be-
tween PIP files e.g. the concept, combining RegionalBusinessTaxIdentifier
and NationalBusinessTaxIdentifier classes, is detected in 3 PIPs (PIP3A5,
PIP3A11 and PIP3B6).
We also performed a manual validation of all detected concepts. We noticed
that among the 90 concepts detected, 79 concepts have effectively a common
semantics. So, we have a detection precision of 87%. From Table 3, we can see
that the value of nosc and ir metrics increase, norc and nolc decrease after
applying our FCA-based approach.
5 Related Work
Ontologies and Formal Concept Analysis (FCA) aim at modeling concepts [1].
For this reason, we use FCA to regroup concepts in a flat ontology representing
general knowledge to improve its inheritance richness. To our knowledge, no
previous work has addressed the problem of flat ontology using FCA techniques.
Some FCA-based proposals in Ontology Engineering differ from our own
strategy with respect to the nature of the problem. Cimiano et al. [1] proposed
a benchmark to discuss how FCA can be used to support Ontology Engineering
and how ontology can be exploited in FCA applications. The FCA can support
the building of the ontology and the constructed ontology can be analyzed and
navigated using FCA techniques [1].
Stumme [12] presents an Ontology Merging approach based on FCA, named
FCA-MERGE, for organizing business knowledge. It takes as input one or more
source ontologies and returns a merged ontology between the given source on-
tologies.
Obitko et al. [9] propose an approach to improve ontology design by discov-
ering the need for new objects (or classes) and relations (properties). They argue
for the necessity of a better description of concepts and relations than just or-
dering them in taxonomy. In our case, we consider instead regrouping concepts
from existing ontology classes for improving the taxonomy representation in a
flat ontology.
6 Conclusion
We have presented an approach for concepts detection in a flat ontology using
a Formal Concept Analysis and applied this methodology on a RosettaNet PIP
Ontology which was automatically generated ontology.
Our goal is to improve the readability, maintainability and inheritance rich-
ness of an ontology. We used FCA to detect groups of concepts sharing properties
and having a common semantics. Through the use of ontology metrics, we have
shown that our FCA-based concept detection methodology can improve inheri-
tance richness.
As the results using RosettaNet are promising, we suggest as future work to
test the efficiency of our concept detection approach to other ontologies dealing
with other domains.
Acknowledgements
We would like to thank RosettaNet Group for allowing us to download the
RosettaNet Partner Interface Processes (PIPs) from the RosettaNet website.
This work has been partially funded by Tunisian Government and NSERC.
References
1. Cimiano, P., Hotho, A., Stumme, G., Tane, J.: Conceptual knowledge processing
with formal concept analysis and ontologies. In: Concept Lattices, pp. 189–207.
Springer (2004)
2. Dogac, A., Kabak, Y., Laleci, G.B.: Enriching ebXML registries with OWL ontolo-
gies for efficient service discovery. Proceedings of the 14th International Workshop
on Research Issues on Data Engineering: Web Services for E-Commerce and E-
Government Applications (RIDE) (2004)
3. Euzenat, J., Shvaiko, P.: Ontology Matching, vol. 18. Springer Heidelberg (2007)
4. Gómez-Pérez, A., Corcho, O., Fernandez-Lopez, M.: Ontological Engineering.
Springer-Verlag, London, Berlin (2002)
5. Jridi, J.E., Lapalme, G.: Adapting RosettaNet B2B standard to Semantic Web
Technologies. vol. 2, pp. 484–491. 15th International Conference on Enterprise
Information Systems, Angers, France (July 2013)
6. Kotinurmi, P., Haller, A., Oren, E.: Ontologically Enhanced RosettaNet B2B In-
tegration. Semantic Web for Business: Cases and Applications (2008)
7. Kotinurmi, P., Haller, A., Oren, E.: Global Business: Concepts, Methodologies,
Tools and Application, vol. 4, chap. Ontologically enhanced RosettaNet B2B In-
tegration, p. 27. USA (2011)
8. Lytras, M., García, R.: Semantic Web applications: a framework for industry and
business exploitation–what is needed for the adoption of the semantic web from
the market and industry. International Journal of Knowledge and Learning 4(1),
93–108 (2008)
9. Obitko, M., Snasel, V., Smid, J., Snasel, V.: Ontology design with formal concept
analysis. Concept Lattices and their Applications, Ostrava: Czech Republic pp.
111–119 (2004)
10. Sheldon, F.T., Jerath, K., Chung, H.: Metrics for maintainability of class inheri-
tance hierarchies. Journal of Software Maintenance and Evolution: Research and
Practice 14(3), 147–160 (2002)
11. Sicilia, M., Rodríguez, D., García-Barriocanal, E., Sanchez-Alonso, S.: Empirical
findings on ontology metrics. Expert Systems with Applications 39(8), 6706–6711
(2012)
12. Stumme, G.: Using ontologies and formal concept analysis for organizing business
knowledge. In: In Proc. Referenzmodellierung 2001. Citeseer (2001)
13. Tartir, S., Arpinar, I.B., Moore, M., Sheth, A.P., Aleman-Meza, B.: OntoQA:
Metric-based ontology quality analysis. In: IEEE Workshop on Knowledge Ac-
quisition from Distributed, Autonomous, Semantically Heterogeneous Data and
Knowledge Sources. vol. 9 (2005)
14. Wang, J., Song, Y.: Architectures supporting RosettaNet. In: Software Engineering
Research, Management and Applications, 2006. Fourth International Conference
on. pp. 31–39. IEEE (2006)