New methods for automated construction of knowledge bases for expert classification systems are being developed using a logic-based language and inductive inference techniques. This research addresses deficiencies of currently available methods for deriving classification rules from empiric data. Unate boolean functions are proposed as a useful language for logical classification rules. An example is given showing the use of an algorithm for converting unate functions to a "criteria table" knowledge representation.
/lp/association-for-computing-machinery/automated-learning-of-rules-for-heuristic-classification-systems-BsPjEMhx0l