Matjaž Šega (2008) BUSINESS USER FRIENDLY RULE EDITING ON THE SEMANTIC WEB. EngD thesis.
The Semantic Web is an evolution of the current Web where information is given semantic meaning. This will enable development of automatic software agents which will carry out various tasks for the users. The implementation of Semantic Web is based on a set of technologies and standards, which form layers. The most important part of the Semantic Web are the two concepts of knowledge representation in the form of ontologies and rules over these ontologies. Although there are already several tools that support ontology development and rules editing, most of them require the users to have expert knowledge of the technologies. A tool called IntelliOnto is being developed that will bring ontology building and rule editing closer to business users. The aim of this graduate thesis was to develop a system for user-friendly rule editing and to implement this system as a rule editing module for IntelliOnto application. The system that was developed enables rules editing without requiring the users to be familiar with abstract syntax. It is based on forms for editing rule atoms. With the help of these forms, any atom of the rule can be presented as a sentence in the natural language. The framework of the form is a text that binds the anchors for arguments. This text is acquired from the ontology where necessary. The text of the argument anchors provides users with information about the type of the argument that is required, while hints of these anchors show additional restrictions when in place. Arguments are entered with the help of the menu with contextually appropriate choices. When the rule is fully edited, it is automatically transformed into the required syntax and sent to a web service. This web service transforms the rule from the abstract into the SWRL syntax and adds it to the ontology. Three approaches to rule editing were developed and implemented. The first one is used for editing one rule at a time in the form of “antecedents => consequents”. The second approach enables rule editing by using a decision table. This approach is especially useful for editing rule packages that contain rules which have bodies consisting of different combinations of condition atoms, entered into decision table. The third approach is based on rule editing in the form of a decision tree, in which branching is done according to validity of the condition atom.
Actions (login required)