Abstract

With the proliferation of the WWW, providing more intelligent Websites has become a major concern in the e-business industry. Recently, this trend has been even more accelerated by the success of Customer Relationship Management (CRM) in terms of product recommendation, and self after service, etc. As a result, many e-companies are eager to embed Web-enabled, rule-based systems, i.e. that is, expert systems, into their Websites, and several well-known commercial tools to facilitate this are already available in the market. So far, most of those tools are based on CGI, but CGI-based systems inherently suffer from problems related to overburdening, when there are too many service demands at the same time. To overcome the limitations of the existing CGI-based expert systems, we propose a new form of Web-enabled expert system that uses a hyperlink-based inference mechanism. In terms of burden to the Web server, our approach has proven to outperform the CGI-based approach theoretically as well as empirically. For practical purposes, our approach is implemented in a software system, WeBIS, and uses a graphic rule editing methodology, Expert's Diagram, that is incorporated into the system to facilitate rule generation and maintenance. WeBIS is now successfully operating for financial consultation on the Website of a leading financial consulting company in Korea.

Keywords: Expert Systems; HTML; Inference; JavaScript; Rule; WWW

Article Outline

1. Introduction

2. Related work

3. Hyperlink-based inference systems

3.1. Hyperlink-based Inference

3.2. FES generation algorithm

3.2.1. Conversion procedure into SNF

3.3. Algorithm extension and its real word example

4. WeBIS approach

4.1. Architecture of WeBIS

4.2. Implementation

5. Performance evaluation

6. Conclusions

References