Information Technology Reference
In-Depth Information
the widely accepted advantage of Java providing portable code and XML providing portable data. In
addition, we use Oracle 9i Database and Application Server platforms (http://www.oracle.com) to imple-
ment the knowledge repository and use the Sun Microsystems Java XML API toolkit to interface the
agents with the XML repository. The decision tree implementation consists of tree nodes with branches
for every category of the node variable. Each traversal from the root node of the decision tree to a leaf
node leads to a separate decision path as illustrated in figure 4. The agents contain methods to traverse
the decision tree and obtain a decision path that can then be translated into an XML object and an XML
document using a DTD file. These form the basis for the generation of decision alternatives and for the
explanations of decisions by the agents. The agents are implemented as java beans and their explana-
tions are available to the user through calls made to their public methods that are exposed as services,
and presented to the user as dynamically generated Web content by using Java Server Pages technology
(http://java.sun.com/products/jsp/index.html).
Business aPPlica tion
Organizations use data mining techniques to leverage the vast amount of data to make better business
decisions (Padmanabhan, et al., 1999; Fan, et al., 2002). Data mining is used for customer profiling in
CRM and customer service support (Hui and Jha, 2000), credit card application approval, fraud detec-
tion, telecommunications network monitoring, market-based analysis (Fayyad et al., 1996), healthcare
quality assurance (Tsechansky et al., 1999) and many other decision-making areas (Brachman et al.,
1996). There is a growing need to not only mine data for decision support, but also to externalize
knowledge from enterprise data warehouses and data marts, to share such knowledge among end us-
ers through automated knowledge discovery and distribution system for effective decision support. In
other words, there is an increasing need for the integration of KMS and DSS systems to meet the needs
of the complex business decision situations. According to Bolloju
et. al.,
“Such integration is expected
to enhance the quality of support provided by the system to decision makers and also to help in build-
ing up organizational memory and knowledge bases. The integration will result in decision support
environments for the next generation,” (Bollojou, 2002, pg. 164). The proposed IKMDSA architecture
illustrates such a next generation integrated KMS and DSS system. The detailed presentation of the
implementation of the architecture is intended to further the research that combines multiple but related
set of research streams such as data mining, automated knowledge discovery, knowledge representation
and storage using XML, knowledge exchange among participating intelligent agents using knowledge
context, and explanation facility (from expert systems research). The authors are currently extending
the architecture in various business domains such as credit approval processing, bankruptcy prediction,
electronic commerce and consumer behavior and Web mining.
Emergent Internet technologies have significant impact on business processes of organizations oper-
ating in the digital economy. Realizing the potential benefits of emergent technologies is dependent on
the effective sharing and use of business intelligence and process knowledge among business partners to
provide accurate, relevant and timely information and knowledge. This requires system models to support
and enable information integration, knowledge exchange and improved collaboration among business
partners. Such systems must provide collaborating partners with intelligent knowledge management
(KM) capabilities for seamless and transparent exchange of dynamic supply and demand information.
Implementing and managing such integration over distributed and heterogeneous information platforms,
Search WWH ::
Custom Search