ABSTRACT
This chapter describes a prototype expert system (ES) for river water pollution determination and prevention. On the ES level, facts in the relational database represent declarative knowledge, or a global database, while relationships embedded in the model enable recognition of elements of the procedural part of the knowledge base. An integral part of the ES is the classification rule, which enables the user of the system to compare the baseline data of the given river pollution levels with the actual results of analysis based on reduced number of measurements. In order to make ESs useful tools to handle problems, a most appropriate knowledge representation must be searched for in a given problem domain, and an effective control of the reasoning process, must be implemented. The overall pollution category of water samples is assigned by an expert panel which includes chemists, biologists and hydrologists.
