XML - based Knowledge Translation Methods for XTT - based Expert Systems ∗

The paper discusses a concept of representing a rule-based knowledge base of an expert system with a hybrid method combining decision trees and tables. The XTT method provides means for logical formulation of a hierarchical rulebase with explicit inference control. It also has a well-defined visual representation, which is useful in the design process. On the machine readable level the XTT knowledge base is encoded in a custom XML format. The paper introduces new methods of automatic XSLT translation of this format into the SVG format. The same method could be used to translate XTT to Prolog code that provides an executable prototype. XSLT is a purely declarative XML-based language for defining document translations. SVG is a upcoming standard for describing vector graphics using XML. Methods presented in the paper provide a purely declarative platform-free translation of XTT knowledge. The translations improve practical design of complex rulebases. They can be implemented in several tool-agnostic methods presented in the paper.