Disjunctive form concept learning system based on genetic algorithm

“Version Space” proposed by Mitchell (1977) is a typical method of concept learning from training examples, but this method has some points to be improved. The purpose of this paper is to construct a flexible learning mechanism which can be applied to critical points. In this paper to do this, the method of concept learning based on genetic algorithms (GA) is proposed. The important features of the algorithm are as follows. Firstly, the system is able to learn the target concept formed by disjunctive normal forms (DNF). Secondly, if there are some incorrect examples in the training examples set, the algorithm will reduce them and generate correct the target concept. This function is called “noise reduction”. Finally, the algorithm is able to learn the target concept from only positive example set