A visual, hierarchical approach to implementing rule-based algorithms in classification of discrete, homogeneous objects

This research examines the graphical approach to implementing an expert system instead of the programming approach of the traditional “if then/else” construct. The rule base and knowledge base are built into the graphical model and only observation on the part of the user is needed for effective discrimination among choices. The application model is designed around a relational database management system with the user interface generated in Visual Basic and a rule base implemented through a hierarchy of mutually exclusive options presented to the user based on choices made as the user progresses through the various decision levels. Decision factors are based on observation of the discriminatory aspects of the objects to be classified as exhibited by the objects and maintained in the knowledge base. Human control is established through explanation and real-time recording and tracking of all choices made for any object under consideration.