Information Technology Reference
In-Depth Information
the provide clear indication of important fields for classification allows the incorporation of explanation
facility, similar to expert systems, among the agents in this type of architecture (Sung et al., 1999).
Moreover, explanation is essential to the interaction between users and knowledge-based systems
(KBS), describing what a system does, how it works, and why its actions are appropriate (Mao and
Benbasat, 2000). Among 87 KBS shell capabilities, users rated explanation facilities and the capability
to customize explanations as the fourth and fifth most important factors, respectively (Stylianou et al.,
1992). Explanation can make KBS conclusions more acceptable (Ye and Johnson, 1995) and builds trust
in a system. The ability of the agents to explain the decision rules used in the decision making process
makes agents powerful tools to aid human agents in complex decision tasks. Such intelligent agent ar-
chitecture, built around well-grounded and well-researched decision models along with standards-based
widely available technologies (such as XML, DTDs), is a significant contribution to furthering research
on agent-based distributed computing and DSS. In the following section, we present the details of IK-
MDSA and discuss its knowledge externalization, knowledge representation, knowledge management
and knowledge delivery mechanism for decision support.
intelligent
knowledge- Based decision suPPort architecture
(IKMdSA)
A KMS has facilities to create, exchange, store and retrieve knowledge in an exchangeable and usable
formatfor decision-making activity. IKMDSA utilizes ID3 algorithms to create knowledge from raw
data to a decision tree representational form. A domain knowledge object represents information about
a specific problem domain in IKMDSA. The domain knowledge object contains information about the
characteristics of the various domain attributes important to the problem domain. The domain knowl-
edge object describes the problem context and provides rules for making decisions in the problem con-
text. The domain knowledge object represents the abstraction used for creating, exchanging and using
modular knowledge objects in IKMDSA. IKMDSA uses intelligent software agents as the mechanism
for encapsulation and exchange of knowledge between agents at the site of knowledge creation and the
Figure 1. Agents have access to Domain Knowledge objects that abstract domain specific knowledge
domain knowledge
agent
+Domain : Object = org.wc DOM
+Rules : Object = org.WC.dom
+DomainName()
+ShareDomainInformation() : Domain Knowledge
+ShareRules() : Domain Knowledge
+ShareKnowledge() : Domain Knowledge
ikMdsa
l earning agent
Knowledge agent
user interface agent
+Agent Knowledge : Domain Knowledge
+Learn() : Domain Knowledge
+Accquire(in DomainKnowledge : Domain Knowledge)
Search WWH ::
Custom Search