A Knowledge-Acquisition Strategy Based on Genetic Programming

In this paper, we have modified our previous GP-based learning strategy to search for an appropriate classification tree. The proposed approach consists of three phases: knowledge creation, knowledge evolution, and knowledge output. One new genetic operator, separation, is designed in the proposed approach to remove contradiction, thus producing more accurate classification rules. A subtree pruning technique is also used to restrain the classification trees excessively expanding in the evolutionary process. Experimental results from diagnosis of breast cancers also show the feasibility of the proposed algorithm.