Restructuring and simplifying rule bases

Rule bases are commonly acquired, by an expert and/or knowledge engineer, in a form which is well-suited for acquisition purposes. When the knowledge base is executed, however, a different structure may be required. Moreover, since human experts normally do not provide the knowledge in compact chunks, rule bases often suffer from redundancy. This may considerably harm their efficiency. In this paper, a procedure is examined to transform rules that are specified in the knowledge acquisition process into an efficient rule base by way of decision tables. This transformation algorithm allows the generation of a minimal rule representation of the knowledge, and verification and optimization of rule bases and other specifications (e.g. legal texts, procedural descriptions). The proposed procedures are fully supported by the PROLOGA tool.