"We are using MINERVA to develop a case-based reasoning (CBR) system that uses Extensible Markup Language (XML) as its case representation language on the World Wide Web." Hideo Shimazu, NEC Corporation

Company:

NEC Corporation www.nec.co.jp telecommunications, computing, electronics

Application:

Case Based Reasoning, Help Desk

Product:

CARET/XML - tool to develop textual CBR applications on the WWW

Reference:

A Textual Case-Based Reasoning System Using XML on the World Wide Web
Hideo Shimazu
EWCBR-98 Sept 1998 Dublin

Abstract:

As Intranet becomes popular, knowledge management systems have become increasingly important in the efforts of corporations to share individuals' experiences. Case-based reasoning provides promising techniques for use in the improvement of such systems.

This paper describes a case-based reasoning (CBR) system that uses Extensible Markup Language (XML) as its case representation language on the World-Wide Web. The paper defines the issues of textual case-based reasoning, then proposes CARET/XML, a textual CBR system architecture.

Cases are written in XML by first-level agents themselves in front offices such as help desks. The architecture carries out similarity-based retrieval using XML tags and values associated with these tags. When a textual case base is constructed by many first-level agents, it is natural that cases are incrementally added their definition.

CARET/XML provides SELECT/OPTION and ANNOTATION tags to define mixed-case representations. A passive case adaptation mechanism is realized in a finer-grained case selection process. The proposed system runs on a commercial relational database management system (RDBMS) and the World-Wide Web.