KNOWLEDGE EXTRACTION GUIDED BY ONTOLOGIES
Database Marketing Application
Filipe Mota Pinto
Politechnic Institute of Leiria, Leiria, Portugal
Teresa Guarda
Superior Institute of Languages and Administration of Leiria, Leiria, Portugal
Keywords: Ontologies, Knowledge Discovery, Databases, Data Mining.
Abstract. The knowledge extraction in large databases has being known as a long term and interactive project. In spite
of such complexity and different options for the knowledge achievement, there is a research opportunity that
could be explored, throughout the ontologies support. Then this support may be used for knowledge sharing
and reuse. This paper describes a research of an ontological approach for leveraging semantic content of
ontologies to improve knowledge extraction in a oil company marketing databases. We attain to analyze
how ontologies and knowledge discovery process may interoperate and present our efforts to propose a
possible framework for a formal integration.
1 INTRODUCTION
At artificial intelligence research area, ontology is
defined as a specification of a conceptualization
(Gruber, 1993). Ontology specifies at a higher level,
the classes of concepts that are relevant to the
domain and the relations that exist between these
classes. Indeed, ontology captures the intrinsic
conceptual structure of a domain. For any given
domain, its ontology forms the heart of the
knowledge representation.
In spite of ontology-engineering tools
development and maturity, ontology integration in
knowledge discovery projects remains almost
unrelated.
Knowledge Discovery in Databases (KDD)
process is comprised of different phases, such as
data selection, preparation, transformation or
modeling. Each one of these phases in the life cycle
might benefit from an ontology-driven approach
which leverages the semantic power of ontologies in
order to fully improve the entire process (Gottgtroy
et al, 2004).
We dare to combine ontological engineering and
KDD process in order to improve it. One of the
promising interests in use of ontologies in KDD
assistance is their use for guiding the process. This
research objective seems to be much more realistic
now that semantic web advances have given rise to
common standards and technologies for expressing
and sharing ontologies (Bernstein et al 2005).
This paper describes an ontological approach
research for leveraging the semantic content of
ontologies to effectively support the knowledge
discovery in databases. We analyze how ontologies
and knowledge discovery process may interoperate
and present our efforts to bridge the two fields,
knowledge discovery in databases and ontology
learning for successful database usage projects.
2 BACKGROUND
There are different relevant topics to the KDD
processes assistance also referred in literature such
as “domain knowledge in KDD” (Pinto and Santos,
2009), “ontology/KDD integration” (Bernstein et al
2005), “KDD life cycle” (Zhou et al., 2009) and
“KDD assisted process” (Bernstein et al 2005).
Ontologies are used to capture knowledge about
some domain of interest. Ontology describes the
concepts in the domain and also the relationships
that hold between those concepts. Different
320
Mota Pinto F. and Guarda T..
KNOWLEDGE EXTRACTION GUIDED BY ONTOLOGIES - Database Marketing Application.
DOI: 10.5220/0003555103200325
In Proceedings of the 13th International Conference on Enterprise Information Systems (ICEIS-2011), pages 320-325
ISBN: 978-989-8425-53-9
Copyright
c
2011 SCITEPRESS (Science and Technology Publications, Lda.)
Figure 1: KDD general phase and task description workflow.
ontology languages provide different facilities.
Ontology Web Language (OWL) is a standard
ontology language from the World Wide Web
Consortium (W3C).
2.1 Semantic Web Language Rule
To the best of our knowledge there are no standard
OWL-based query languages. Several RDF -based
query languages exist but they do not capture the full
semantic richness of OWL. To tackle this problem,
it was developed a set of built-in libraries for
Semantic Web Rule Language (SWRL) that allow it
to be used as a query language
The OWL is a very useful means for capturing
the basic classes and properties relevant to a domain.
However, these domain ontologies establish a
language of discourse for eliciting more complex
domain knowledge from subject specialists. Due to
the nature of OWL, these more complex knowledge
structures are either not easily represented in OWL
or, in many cases, are not representable in OWL at
all. The classic example of such a case is the
relationship uncleOf(X,Y). This relation, and many
others like it, requires the ability to constrain the
value of a property (brotherOf) of one term (X) to be
the value of a property (childOf) of the other term
(Y); in other words, the siblingOf property applied
to X (i.e., brotherOf(X,Z)) must produce a result Z
that is also a value of the childOf property when
applied to Y (i.e., childOf(Y,Z)). This “joining” of
relations is outside of the representation power of
OWL.
One way to represent knowledge requiring joins
of this sort is through the use of the implication (Æ)
and conjunction (AND) operators found in rule-
based languages (e.g., SWRL). The rule for the
uncleOf relationship appears as follows:
brotherOf(X,Z) AND
childOf(Y,Z)→uncleOf(X,Y)
2.2 Evaluation of Knowledge Reuse
Effectiveness
The main objective of this research is to assist the
KDD process based on ontology knowledge.
Therefore, it is assumed that effectively ontology
has learned from KDD domain and practice. Thus it
is possible to provide users with information that are
relevant to their needs at each of KDD phases.
Hence, the related task (process option)
suggestion returned by the ontology will be the
primary basis to determine the quality of the relevant
information retrieved. For the present purpose we
admit as major indices, precision and recall (Han
and Kamber, 2001):
Precision=(Relevant
∩Selected)/(Selected_Results
)
Precision expresses the proportion of related
results (Relevant∩Selected) among relevant results
retrieved (Selected_Results). In other words, to
reflect the amount of knowledge correctly identified
(in the ontology) with respect to the whole
knowledge available in the ontology As related
results we intend the entire set of ontology elements
(classes and data properties) related to the subject
(e.g., to preprocessing phase: set of related classes
and relationships). Also, we use selected results as
the set of related results and selected at the user
Data
Target Data
PréProcessed
Data
Transformed
Data
Patterns
Knowledge
Pré-
Processing
Transformation
Data
Mining
Interpretation
Evaluation
Selection
Objectives
PreProcessing
Modeling
Evaluation
Domain Objectives
Research scope
Problem Perception
Data requirements definition
Data understand
Pre-processing tasks:
- dictionary
- input/output process
- data requirements
Transformation
Transformation tasks:
- dictionary
- input/output process
- data requirements
Modeling dictionary:
-tasks
- methods
- input specifications
- output models
- business results type
Models Evaluation:
- evaluation type dictionary
Business models deployment:
- deployment approach dictionary
KNOWLEDGE EXTRACTION GUIDED BY ONTOLOGIES - Database Marketing Application
321
question (e.g., set of suggested results for, e.g.,
birthDate attribute preprocessing).
Recall=(Relevant
∩Selected)/(Relevant_Results)
Recall expresses the proportion of results
retrieved (Relevant∩Selected) from related results
(Relevant_Results). It is used to reflect the amount
of knowledge correctly identified with respect to all
the knowledge that it should identified.
During this work, precision will be used to
evaluate the proportion of user interests towards the
KDD phase assistance. This proportion examines
how correct the ontology is suggesting tasks
(options) when solicited by the user. On other hand,
recall, estimates the ability that the system is able to
satisfy user needs.
2.3 Database Marketing
Nowadays, organizations try to act dynamically in
competitive markets so that they can find and keep
their customers. In this context Database Marketing
presents itself as a privileged tool marketing
professionals may use to do so.
DBM refers the use of database technology for
supporting marketing activities (Shepard 1998).
Therefore DBM is a marketing process driven by
information and managed by database technology
(Grassl 2007), allowing marketing professionals to
develop and to implement better marketing
programs and strategies. However, DBM is usually
approached using classical statistical inference,
which may fail when complex, multi-dimensional,
and incomplete data is available.
Through the advances in information and
communication technologies, corporations can
effectively obtain and store transactional and
demographic data on individual customers at
reasonable costs. The DBM activity framework has
changed significantly over the last years. In past,
database marketers applied business rules to target
customers directly, based sometimes in the
marketer’s intuition. The current approach relies on
predictive response models to target customers for
offers or other marketing purposes. The challenge
now is how to extract important knowledge from
those vast databases. Through the use of DBM
organizations get to understand the customers’
preferences and behaviours through analyzing their
transactional data. However, in almost cases, the
data set presents several problems as the result of
procedural factors, inadequate questionnaire options,
refusal of response, or/and inadequate database
schema's.
The set of DBM processes used in this study
include the use of a KDD framework in order to
create predictive models. These models may be used
to, e.g., accurately estimate the probability that a
customer will respond to a specific offer and can
significantly increase the response rate to a product
offering. The old model of design-build-sell is now
being replaced by sell-build-redesign.
3 RESEARCH APPROACH
Here, we attain to the effective support the KDD
process using ontologies. In order to do this we have
collected a marketing database from a multinational
oil company. Therefore, we used a real case study in
order to effectively test and deploy to ontological
work in the KDD process (Pinto et al, 2009). Our
approach is defined in two distinct steps:
- Marketing database collection and
preparation;
- Practical KDD ontology based
development;
3.1 Marketing Database
One of the most important marketing tools used by
oil companies for customer fidelization is the
marketing card loyalty programs (Phillips. and
Buchanan, 2001). This approach allows cardholders
to obtain fuel purchase discounts, to participate in
marketing campaigns or to become members of a
restrict club with restrict privileges.
Since it is an open marketing system program
where all oil customers may access, from the
company perspective this will turn into an important
information source for almost customer oriented
marketing strategies definition or product offer
policies.
We have collected a card loyalty program
marketing database from a multinational company.
This database has three main tables: card owner,
station and transactions. The available data refers to
the past two year’s activity. The data structure is as
follows:
ICEIS 2011 - 13th International Conference on Enterprise Information Systems
322
Table 1: Data Table Card Owner.
Field Description Type Domainrange
IdCard cardidentification Primarykey
Idclient Clientidentification Foreignerkey
birthDate Clientbirthdate Date <today
cardClientDate Startingcardownerdate Date <today
cardInitialDate Startingclientdate Date <today
Postcode Zipcode Integer <10000
postCod3 3Zipcode Integer <1000
maritalStatus Maritalstatus String {cas;sol;div;viv;out}
Gender Clientsex String {mas,fem}
vehicleType Vehicletypedescription String {lig,merc,pes,out}
vehicleYear Vehicleidentificationyeardate Number <10000
fuelType Fueldescriptiontype String {diesel,gasolina,gpl,out}
Table 2: Data Table Card Transactions.
Field Description Type Domainrange
IdMov Transactionidentification Primarykey
IdCard Cardidentification Foreignerkey
Date Transactiondate Date <today
fuelValue Fueltransactionamount real <10000
fuelLitres Transactionlitersamount real <3000
shopValue Transactionvalueamount real <10000
shopUnits Shopunitstransactionamount integer <10
stationCode Fuelstationidentification Foreignerkey
Table 3: Data Table Station.
Field Description Type Domainrange
stationCode Fuelstationidentificationcode Primarykey
stationType Stationidentificationtype String {urb,rur,est}
postCode Zipcode integer <10000
postCod3 3Zipcode integer <1000
4 PRACTICAL KDD
ONTOLOGY BASED
DEVELOPMENT
For the KDD development we have based our work
on the free open source WEKA toolkit (Witten and
Frank2000). For ontological support we have used
the PROTÉGÉ OWL editor and SWRL language
(Knublauch et al. 2005).
Since KDD process generates output models, it
was considered useful to represent them in a
computable way. Such representation works as a
general description of all options taken during the
process. Based on PMML descriptive DM model we
have introduced an OWL class in our ontology
named ResultModel which holds instances with
general form:
ResultModel {
domain Objective Type;
algorithm;algorithmTasks;
algorithmParameters;
workingAlgorithmDataSet;
EvaluationValue;
DeploymentValue}
Since the ontology contribution to the KDD
process is quantitatively uncertain we have used a
quality approach based on KDD team individual
expert contribution.
4.1 Ontological Work
One of the promising interest of ontologies is they
common understand for sharing and reuse. Hence
we have explored this characteristic to effectively
assist the KDD process. Indeed, this research
presented the KDD assistance at two levels: Overall
KNOWLEDGE EXTRACTION GUIDED BY ONTOLOGIES - Database Marketing Application
323
process assistance based on ModelResult class and,
- KDD phase assistance. Since our ontology has
a formal structure related to KDD process, is able to
infer some result at each phase.
To this end, user need to invoke the system rule
engine (reasoner) indicating some relevant
information, e.g., at data preprocessing task:
swrl:query hasDataPreprocessingTask(?dpp,”ds”),
where hasDataPreProcessingTask is an OWL
property which infers from ontology all assigned
data type preprocessing tasks (dpp) related to each
attribute type within the data set “ds”. Moreover,
user is also assisted in terms of ontology capability
index, through the ontology index - precision, recall
and PRI metrics.
Once we have a set of running KDD process
registered at the knowledge base, whenever a new
KDD process starts one the ontology may support
the user at different KDD phases. As example to a
new classification process execution the user
interaction with ontology will follow the framework
as described in next section. The ontology will lead
user efforts towards the knowledge extraction
suggesting by context. That is the ontology will act
accordingly to user question, e.g., at domain
objective definition (presented by user) the ontology
will infer which is type of objectives does the
ontology has. All inference work is dependent of
previous loaded knowledge. Hence, there is an
ontology limitation – only may assist in KDD
process which has some similar characteristics to
others already registered.
5 EXPERIMENT
To build up mining experiments we have used Weka
Toolkit (Witten and Frank, 2000) which allowed not
only the actual mining but also featured analysis and
algorithm evaluation. These experiments did not aim
to the full construction of a classification model but
instead to test and analyze different approaches and
further ranking.
Our system prototype operation follows general
KDD framework (Figure 1) and uses the ontology to
assist at each user interaction. Our experimentation
was developed over a real oil company fidelity card
marketing database. This database has three main
tables: card owner; card transactions and fuel
station.
To carry out this we have developed an initial set
of SWRL rules. Since KDD is an interactive
process, these rules deal at both levels: user and
ontological levels. The logic captured by these rules
is this section using an abstract SWRL
representation, in which variables are prefaced with
question marks.
Domain objective: customer profile
Modeling objective: description
Initial database: fuel fidelity card;
Database structure: 4 tables;
The most relevant rule extracted from above data
algorithms use was:
if (age<27 and vehicleType=”Lig” and
sex=”Female”) then 1stUsed=”p”
In this model we may say that, female card owners
under 27 years of age have a “lig” (ligeiro) category
car and use a fuel station located in range of 10
kilometers from their address.
Also, practical KDD process tasks have been
done supported by SWRL ontology queries. This
query tasks was manually performed by the user.
Therefore, the guidance was accomplished and
achieved throughout knowledge base updating with
the general model:
INSERT record KNOWLEDGE BASE
hasAlgorithm(J48) AND
hasModelingObjectiveType(classif
ication) AND
hasAlgorithmWorkingData
({idCard; age; carClientGap;
civilStatus; sex; vehicleType;
vehicleAge;nTransactions;
tLiters; tAmountFuel;tQtdShop;
1stUsed; 2stUsed; 3stUsed }) AND
Evaluation(67,41%; 95,5%) AND
hasResultMoldel (J48;
classification;“wds”,PCC;0,674;0
955)
The evaluation, once performed, the system
automatically updates the knowledge base with a
new record. The registered information will serve
for future use – knowledge sharing and reuse.
From the aforementioned previous research work
(Pinto and Santos, 2009), we also have used the
output models and integrated them into the
knowledge base.
6 CONCLUSIONS AND
FURTHER RESEARCH
The KDD success still is very much user dependent.
Though our system may suggest a valid set of tasks
which better fits in KDD process design, it still miss
the capability of automatically runs the data, develop
modeling approaches and apply algorithms.
This work strived to improve KDD process
supported by ontologies. To this end, we have used
ICEIS 2011 - 13th International Conference on Enterprise Information Systems
324
general domain ontology to assist the knowledge
extraction from databases with KDD process.
This research focuses the KDD development
assisted by ontologies. Moreover we use ontologies
to simplify and structure the development of
knowledge discovery applications offering to a
domain expert a reference model for the different
kind of DM tasks, methodologies to solve a given
problem, and helping to find the appropriate
solution.
Future research work will be devoted to expand
the use of KDD ontology through knowledge base
population with more relevant concepts about the
process. Another interesting direction to investigate
is to represent the whole knowledge base in order to
allow its automatic reuse.
REFERENCES
Bernstein, A., Provost, F., and Hill, S. (2005). Toward
intelligent assistance for a data mining process. IEEE
Transactions knowledge & data engineering, 17(4).
Domingos, P. (2003). Prospects and challenges for multi-
relational data mining. SIGKDD Explorer Newsletter,
5(1):80–83.
Pinto, Filipe. Mota, Gago, P., and Santos, M. F. (2009).
Marketing database knowledge extraction In IEEE
13th International Conference on Intelligent
Engineering Systems 2009.
Pinto, Mota Filipe. and Santos, M. F. (2009). Database
marketing supported by ontologies . In 11th ICEIS.
Gottgtroy, P., Kasabov, N., and MacDonell, S. (2004). An
ontology driven approach for knowledge discovery in
biomedicine.
Grassl Wolfgang (2007) “The reality of brands: Towards
an ontology of marketing. American Journal of
Economics and Sociology, 58(2):313–319, April.
Gruber, T. R. (1993). A translation approach ontology
specifications. Knowledge Acquisition, 5:199–220.
Knublauch Holger, Horridge M., Noy N, and Wang H
(2005) “The Protégé OWL Experience”, Workshop on
OWL: Experiences & Directions, Galway, Ireland
Han, J. and Kamber, M. (2001). Data mining: concepts &
Techniques. Morgan Kaufman, San Francisco, CA.
Horrocks, I., Patel-Schneider, S., Grosof, B., and Dean, M.
(2004). Swrl: A semantic web rule language -
combining owl and ruleml. Technical report, W3C.
Phillips, J. and Buchanan, B. G. (2001). Ontology-guided
knowledge discovery in databases. In ACM, editor,
International Conference On Knowledge Capture , p
123–130. I C. On Knowledge Capture.
Sarwar, B., Karypis, G., Konstan, J., and Riedl, J. (2000).
Analysis of recommendation algorithms for e-
commerce. In Proc 2
nd
ACM Conf e-Commerce.
Shepard David. (1998)“Database Marketing”. São Paulo:
Makron Books.
Witten, I. H. and Frank, E. (2000). Data Mining: Practical
Machine Learning Tools and Technique. The Morgan
Kaufmann Series in Data Management Systems, 2nd
edition.
Zhou Xuan, Geller James, and Perl Yehoshua and Halper
Michael.(2009) “An Application Intersection
Marketing Ontology”, Theoretical Computer Science,
pp 143–163.LNCS. Springer Berlin.
KNOWLEDGE EXTRACTION GUIDED BY ONTOLOGIES - Database Marketing Application
325