Information Technology Reference
In-Depth Information
The explanatory power afforded by decision trees comes from generation of understandable rules,
clear identification of fields that are most important for prediction and classification and the incorporation
of explanation facility. 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). Explanation can make KBS conclusions more acceptable (Ye and Johnson, 1995)
and build trust in a system (Swartout, 1983). Decision trees lend themselves to automatic generation of
structured queries to extract pertinent data from organizational data repositories making them particularly
useful in providing insights and explanations for the non-technical user (Apte and Weiss, 1997). Deci-
sion trees are especially suitable for decision problems that require generation of human understandable
decision rules based on a mix of classification of categorical and continuous data (Quinlan, 1996). They
clearly indicate the importance of individual data fields to the decision problem and reduce the cognitive
burden of the decision maker (Mao and Benbasat, 2000). Decision trees represent a powerful and easily
interpretable technique for modeling business decisions that can be reduced to a rule-based form.
XMl and decision t rees for knowledge
r ePresent ation and
eXchange
xML and document Type definitions (dTds)
HTML provides a fixed set of tags that are used to markup content (information) primarily for consump-
tion by human beings. Despite its efficiency for presenting information in human readable format, HTML
is very limited in extensibility and customization of markup tags and description of the data contained
in those tags. This constraint limits the use of HTML by application software for information sharing
in a distributed computing environment where application programs, including intelligent agents are
expected to work with available data, rules and knowledge without human intervention.
The use of XML and its related set of standards, developed by the World Wide Web Consortium,
(W3C http://www.w3c.org), have helped overcome some of these limitations. XML allows for the cre-
ation of custom tags that contain data from specific domains. XML is a meta-language that allows for
the creation of languages that can be represented by customized XML tags. For example, a company
in the furniture industry may develop customized tags for the representation of content to serve its
business domain. By creating custom tags, the company can represent the data in a more meaningful
and flexible way than it could using HTML. The company may also develop documents that represent
business-rules using XML that can shared either with human agents or with software agents.Unambigu-
ous understanding of the content of customized XML tags by interested parties requires description of
both the content and structure of XML documents. This description of structures in XML documents is
provided by the XML schema which can be written following the set of standards called XML Schema
and/or the document type definition (DTD) language as adopted and standardized by the W3C. XML
schema describes specific elements, their relationships and specific types of data that can be stored in
each tag. XML documents can be validated and parsed by application software provided either the DTD
or the XML Schema of the corresponding document is made available. XML parsers written in C, C++
or Java can process and validate XML documents (containing business rules and data) based on XML
schemas written based on either the DTD or the XML schema specification. Application software ap-
propriate parser utilities are able to read and/or write to XML documents following the W3C standards
Search WWH ::
Custom Search