Programming model for concept learning and its solution based on genetic algorithms

Learning from examples is an important branch of inductive learning, and is also the bottleneck in concepts extraction of machine learning. Based on inductive learning theory this paper applies combinatorial optimization method to setup the programming models of learning concepts of the prepositional logic formulas in the conjunctive normal form (CNF) and disjunctive normal form (DNF). Then, genetic algorithms (GA), specified to CNF learning, is designed. GA can find the multiple optimal solution in theory and practice, and experiments reveal that it runs more efficiently compared with heuristic algorithms of the generalisation-and-specialisation (GS) type.