Extention of Learning by Switching Rule Generation and Reasoning Methods ---Learning Various Data Set under Switching-Timing Rules Previously Acquired

Abstract

When we solve a problem, we firstly have no knowledge and gradually acquire some piece of knowledge by observing new data, and at last arrive at complete knowledge for solving the problem. To implement such kind of learning mechanism, we proposed a learning method of switching reasoning methods and rule generation methods. In the previous paper, we proposed a method that aquires meta-rules for switching with a reinforcement learning. The aquired meta-rules may be independent from data sets and can be commonly used for learning many data sets. In this paper, we apply the switching meta-rules acquired from the iris plants data set to learning the wine recognition data set. We got a result that has the smaller number of rules than the conventional switching method with the same correct ratio.