Description
An Ontology-based Business Intelligence Application In A Financial Knowledge Management System
An ontology-based business intelligence application in a ?nancial knowledge
management system
Hilary Cheng
a,
*
, Yi-Chuan Lu
b,1
, Calvin Sheu
c,2
a
Department of Business Administration, Yuan Ze University, Chung-Li 320, Taiwan
b
Department of Information Management, Yuan Ze University, Chung-Li 320, Taiwan
c
Graduate School of Management, Yuan Ze University, Chung-Li 320, Taiwan
a r t i c l e i n f o
Keywords:
Business intelligence
Data mining
Knowledge management
Bond ratings
a b s t r a c t
Business intelligence (BI) applications within an enterprise range over enterprise reporting, cube and ad
hoc query analysis, statistical analysis, data mining, and proactive report delivery and alerting. The most
sophisticated applications of BI are statistical analysis and data mining, which involve mathematical and
statistical treatment of data for correlation analysis, trend analysis, hypothesis testing, and predictive
analysis. They are used by relatively small groups of users consisting of information analysts and power
users, for whom data and analysis are their primary jobs. We present an ontology-based approach for BI
applications, speci?cally in statistical analysis and data mining. We implemented our approach in ?nan-
cial knowledge management system (FKMS), which is able to do: (i) data extraction, transformation and
loading, (ii) data cubes creation and retrieval, (iii) statistical analysis and data mining, (iv) experiment
metadata management, (v) experiment retrieval for new problem solving. The resulting knowledge from
each experiment de?ned as a knowledge set consisting of strings of data, model, parameters, and reports
are stored, shared, disseminated, and thus helpful to support decision making. We ?nally illustrate the
above claims with a process of applying data mining techniques to support corporate bonds classi?cation.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
Knowledge is power. Today’s business environment has been
tougher than ever. Enterprises experience global competitions.
Customers demand more on product features and services. Corpo-
rate expenses are continuously increasing. To survive in the harsh
environment, high-level management needs business intelligent
information to ef?ciently manage corporate operations and sup-
port their making of decisions. Support-level staffs need knowl-
edge information to provide better customer services for gaining
satisfaction and retaining loyalty. Vast operating data is staggered
into various corporate databases and needs consolidating. It has
become more important than ever to access and generate valuable
knowledge and share information among authorized users within a
corporation and/or business partners. Thus, a system of integrating
knowledge management and decision support processes is in great
demand. As mentioned in (Bolloju, Khalifa, & Turban, 2002), a syn-
ergy can be created by the integration of decision support and
knowledge management, since these two processes involve activi-
ties that complement each other. The knowledge retrieval, storage,
and dissemination activities in knowledge management function-
ality enhance the dynamic creation and maintenance of decision
support models, subsequently, enhancing the decision support
process. From the system design’s point of view, what we need is
a new generation of knowledge-enabled system that provides
enterprise an infrastructure to capture, cleanse, store, organize,
leverage, and disseminate not only source data and information
but also the knowledge or value-added information of the ?rm
(Nemati, Steiger, Iyer, & Herschel, 2002).
We present the concept of ?nancial knowledge management
system (FKMS), which is a prototype of KM environment speci?-
cally for ?nancial research purposes. What the environment gen-
erates is groups of knowledge set with strings of data, models,
parameters, and reports. Ontology of knowledge management
and knowledge sharing is presented. Finally, a realization of deci-
sion support and knowledge sharing processes to a corporate
bond classi?cation is illustrated. With FKMS, knowledge workers
can freely extract sets of ?nancial and economic data, analyze
data with different decision support modules, rerun experi-
ments with different sets of parameters, and ?nally disseminate
value-added information (knowledge) through middleware or
Internet to remote clients. Not to mention that the knowledge
0957-4174/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2008.02.047
* Corresponding author. Tel.: +886 3 4638800x2626; fax: +886 3 4630377.
E-mail addresses: [email protected] (H. Cheng), [email protected].
edu.tw (Y.-C. Lu), [email protected] (C. Sheu).
1
Tel.: +886 3 4638800x2600; fax: +886 3 4352077.
2
Tel.: +886 3 4638800x2693; fax: +886 3 4630377.
Expert Systems with Applications 36 (2009) 3614–3622
Contents lists available at ScienceDirect
Expert Systems with Applications
j our nal homepage: www. el sevi er . com/ l ocat e/ eswa
generated is being collected, classi?ed, and shared with col-
leagues, and thus well archived into corporate business intelli-
gence databank.
The remainder of this paper proceeds as follows. Section 2 rea-
sons our motivation for developing FKMS. Section 3 introduces sys-
tem architecture of FKMS. Section 4 presents the ontology of
knowledge management and knowledge sharing, and demonstrates
with a case of corporate bond classi?cation problem. Section 5
concludes this paper.
2. Motivation of developing a ?nancial knowledge
management system
One of the biggest challenges that most security investment
institutions experienced was the lack of an intelligent data mining
system to support investment researches decisions. The problems
that their system encountered including:
(i) lack of ef?ciency in managing vast ?nancial data,
(ii) lack of communication and knowledge sharing among
analysts,
(iii) lack of a mechanism to resolve synchronization problems
when multiple users are accessing data,
(iv) lack of a mechanism to ef?ciently manage generated
research experiments,
(v) lack of an automation to publish its reports to clients via its
Web sites or email.
Though data for ?nancial applications are simple data, the data
typically includes time series information and the relationships
among the ?nancial instruments are complex. For example, con-
sider a derivative security objects. The derivative security object
often shares underlying securities with other derivatives. Underly-
ing securities can come from many classes of instruments, from a
simple currency to an interest rate swap to a hedge. As the securi-
ties become more complex, the problems of data management and
knowledge discovery become more dif?cult. Consider a security
portfolio, the portfolio construction is a process of quantitative
analysis over massive amounts of data. The data cube and ad hoc
analysis techniques are an invisible solution to support this pro-
cess. A system that ef?ciently supports ?nancial application thus
would provide support for: (i) temporal objects, (ii) object manage-
ment, which is the ef?cient storage and manipulation of complex
data, (iii) knowledge discovery, the capability of extracting infor-
mation as rules for decision making. The (i) and (ii) result in ?nan-
cial data modeling using object-oriented techniques, whereas the
(iii) is merely data mining techniques.
3. System architecture
The architecture of the FKMS is a layered structure as shown in
Fig. 1. The object-oriented design gives the system ?exibility and
expandability. FKMS consists of ?ve layers: the resource layer,
the data conversion layer, the data storage and management layer,
the knowledge/trend/pattern layer, and ?nally the user process
layer. Various sources of data are converted into the time series
database according to prede?ned schema. With OLAP tool, users
can easily de?ne various periodical reports with report generator
and generate sets of data cubes for analysis. The resulting data
cubes are stored and managed in the FKMS that various valuation
models, data mining techniques or statistical modules can be ap-
plied to. In addition, Web pages represented by the XML can be
sent to major corporate clients (as a message), as well as posted
on the enterprise information portal with Web-enabled modules
and messaging tools. Finally, thousands of ?les generated by the
analysts are well managed and monitored for knowledge sharing
and as for internal performance evaluations.
The data cubes are stored in a traditional relational database
management system (RDBMS); users can easily divert the data
cubes via ODBC or JDBC for analytical applications at the knowl-
edge/trend/pattern layer. The selected analytical applications are
either designed or programmed by users, or the off-the-shell soft-
ware such as Excel, Matlab, IMSL, SAS, SPSS, or S-Plus. A use case
diagram in Fig. 2 depicted the function requirement for FKMS
implementation.
3.1. Resource layer
Various resources of data are used by analysts when they write
research reports or run valuation models. Typical examples of
these data resources are ?nancial databases from foreign data ven-
dors such as Bloomberg, Data Stream, First Call, or from domestic
data vendors like the TEJ and the SFI, as well as other reports from
competitors, and some periodically published data on the Web
sites. While some data are static, meaning that they are periodi-
cally released, some are dynamic, which means that they are not
periodic. On the other hand, the format of data can be classi?ed
as structured, semi-structured, and unstructured. We focus on
the management of structured data, as their format is clearly de-
?ned so that data operations or manipulations can be deployed.
3.2. Data conversion layer
A data warehouse should always provide its users with accu-
rate, consistent, and real-time data. It should be ?exible to support
all corporate operations and changes. Corporations usually manage
Fig. 1. The system architecture of ?nancial knowledge management system.
H. Cheng et al. / Expert Systems with Applications 36 (2009) 3614–3622 3615
doc_281469647.pdf
An Ontology-based Business Intelligence Application In A Financial Knowledge Management System
An ontology-based business intelligence application in a ?nancial knowledge
management system
Hilary Cheng
a,
*
, Yi-Chuan Lu
b,1
, Calvin Sheu
c,2
a
Department of Business Administration, Yuan Ze University, Chung-Li 320, Taiwan
b
Department of Information Management, Yuan Ze University, Chung-Li 320, Taiwan
c
Graduate School of Management, Yuan Ze University, Chung-Li 320, Taiwan
a r t i c l e i n f o
Keywords:
Business intelligence
Data mining
Knowledge management
Bond ratings
a b s t r a c t
Business intelligence (BI) applications within an enterprise range over enterprise reporting, cube and ad
hoc query analysis, statistical analysis, data mining, and proactive report delivery and alerting. The most
sophisticated applications of BI are statistical analysis and data mining, which involve mathematical and
statistical treatment of data for correlation analysis, trend analysis, hypothesis testing, and predictive
analysis. They are used by relatively small groups of users consisting of information analysts and power
users, for whom data and analysis are their primary jobs. We present an ontology-based approach for BI
applications, speci?cally in statistical analysis and data mining. We implemented our approach in ?nan-
cial knowledge management system (FKMS), which is able to do: (i) data extraction, transformation and
loading, (ii) data cubes creation and retrieval, (iii) statistical analysis and data mining, (iv) experiment
metadata management, (v) experiment retrieval for new problem solving. The resulting knowledge from
each experiment de?ned as a knowledge set consisting of strings of data, model, parameters, and reports
are stored, shared, disseminated, and thus helpful to support decision making. We ?nally illustrate the
above claims with a process of applying data mining techniques to support corporate bonds classi?cation.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
Knowledge is power. Today’s business environment has been
tougher than ever. Enterprises experience global competitions.
Customers demand more on product features and services. Corpo-
rate expenses are continuously increasing. To survive in the harsh
environment, high-level management needs business intelligent
information to ef?ciently manage corporate operations and sup-
port their making of decisions. Support-level staffs need knowl-
edge information to provide better customer services for gaining
satisfaction and retaining loyalty. Vast operating data is staggered
into various corporate databases and needs consolidating. It has
become more important than ever to access and generate valuable
knowledge and share information among authorized users within a
corporation and/or business partners. Thus, a system of integrating
knowledge management and decision support processes is in great
demand. As mentioned in (Bolloju, Khalifa, & Turban, 2002), a syn-
ergy can be created by the integration of decision support and
knowledge management, since these two processes involve activi-
ties that complement each other. The knowledge retrieval, storage,
and dissemination activities in knowledge management function-
ality enhance the dynamic creation and maintenance of decision
support models, subsequently, enhancing the decision support
process. From the system design’s point of view, what we need is
a new generation of knowledge-enabled system that provides
enterprise an infrastructure to capture, cleanse, store, organize,
leverage, and disseminate not only source data and information
but also the knowledge or value-added information of the ?rm
(Nemati, Steiger, Iyer, & Herschel, 2002).
We present the concept of ?nancial knowledge management
system (FKMS), which is a prototype of KM environment speci?-
cally for ?nancial research purposes. What the environment gen-
erates is groups of knowledge set with strings of data, models,
parameters, and reports. Ontology of knowledge management
and knowledge sharing is presented. Finally, a realization of deci-
sion support and knowledge sharing processes to a corporate
bond classi?cation is illustrated. With FKMS, knowledge workers
can freely extract sets of ?nancial and economic data, analyze
data with different decision support modules, rerun experi-
ments with different sets of parameters, and ?nally disseminate
value-added information (knowledge) through middleware or
Internet to remote clients. Not to mention that the knowledge
0957-4174/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2008.02.047
* Corresponding author. Tel.: +886 3 4638800x2626; fax: +886 3 4630377.
E-mail addresses: [email protected] (H. Cheng), [email protected].
edu.tw (Y.-C. Lu), [email protected] (C. Sheu).
1
Tel.: +886 3 4638800x2600; fax: +886 3 4352077.
2
Tel.: +886 3 4638800x2693; fax: +886 3 4630377.
Expert Systems with Applications 36 (2009) 3614–3622
Contents lists available at ScienceDirect
Expert Systems with Applications
j our nal homepage: www. el sevi er . com/ l ocat e/ eswa
generated is being collected, classi?ed, and shared with col-
leagues, and thus well archived into corporate business intelli-
gence databank.
The remainder of this paper proceeds as follows. Section 2 rea-
sons our motivation for developing FKMS. Section 3 introduces sys-
tem architecture of FKMS. Section 4 presents the ontology of
knowledge management and knowledge sharing, and demonstrates
with a case of corporate bond classi?cation problem. Section 5
concludes this paper.
2. Motivation of developing a ?nancial knowledge
management system
One of the biggest challenges that most security investment
institutions experienced was the lack of an intelligent data mining
system to support investment researches decisions. The problems
that their system encountered including:
(i) lack of ef?ciency in managing vast ?nancial data,
(ii) lack of communication and knowledge sharing among
analysts,
(iii) lack of a mechanism to resolve synchronization problems
when multiple users are accessing data,
(iv) lack of a mechanism to ef?ciently manage generated
research experiments,
(v) lack of an automation to publish its reports to clients via its
Web sites or email.
Though data for ?nancial applications are simple data, the data
typically includes time series information and the relationships
among the ?nancial instruments are complex. For example, con-
sider a derivative security objects. The derivative security object
often shares underlying securities with other derivatives. Underly-
ing securities can come from many classes of instruments, from a
simple currency to an interest rate swap to a hedge. As the securi-
ties become more complex, the problems of data management and
knowledge discovery become more dif?cult. Consider a security
portfolio, the portfolio construction is a process of quantitative
analysis over massive amounts of data. The data cube and ad hoc
analysis techniques are an invisible solution to support this pro-
cess. A system that ef?ciently supports ?nancial application thus
would provide support for: (i) temporal objects, (ii) object manage-
ment, which is the ef?cient storage and manipulation of complex
data, (iii) knowledge discovery, the capability of extracting infor-
mation as rules for decision making. The (i) and (ii) result in ?nan-
cial data modeling using object-oriented techniques, whereas the
(iii) is merely data mining techniques.
3. System architecture
The architecture of the FKMS is a layered structure as shown in
Fig. 1. The object-oriented design gives the system ?exibility and
expandability. FKMS consists of ?ve layers: the resource layer,
the data conversion layer, the data storage and management layer,
the knowledge/trend/pattern layer, and ?nally the user process
layer. Various sources of data are converted into the time series
database according to prede?ned schema. With OLAP tool, users
can easily de?ne various periodical reports with report generator
and generate sets of data cubes for analysis. The resulting data
cubes are stored and managed in the FKMS that various valuation
models, data mining techniques or statistical modules can be ap-
plied to. In addition, Web pages represented by the XML can be
sent to major corporate clients (as a message), as well as posted
on the enterprise information portal with Web-enabled modules
and messaging tools. Finally, thousands of ?les generated by the
analysts are well managed and monitored for knowledge sharing
and as for internal performance evaluations.
The data cubes are stored in a traditional relational database
management system (RDBMS); users can easily divert the data
cubes via ODBC or JDBC for analytical applications at the knowl-
edge/trend/pattern layer. The selected analytical applications are
either designed or programmed by users, or the off-the-shell soft-
ware such as Excel, Matlab, IMSL, SAS, SPSS, or S-Plus. A use case
diagram in Fig. 2 depicted the function requirement for FKMS
implementation.
3.1. Resource layer
Various resources of data are used by analysts when they write
research reports or run valuation models. Typical examples of
these data resources are ?nancial databases from foreign data ven-
dors such as Bloomberg, Data Stream, First Call, or from domestic
data vendors like the TEJ and the SFI, as well as other reports from
competitors, and some periodically published data on the Web
sites. While some data are static, meaning that they are periodi-
cally released, some are dynamic, which means that they are not
periodic. On the other hand, the format of data can be classi?ed
as structured, semi-structured, and unstructured. We focus on
the management of structured data, as their format is clearly de-
?ned so that data operations or manipulations can be deployed.
3.2. Data conversion layer
A data warehouse should always provide its users with accu-
rate, consistent, and real-time data. It should be ?exible to support
all corporate operations and changes. Corporations usually manage
Fig. 1. The system architecture of ?nancial knowledge management system.
H. Cheng et al. / Expert Systems with Applications 36 (2009) 3614–3622 3615
doc_281469647.pdf