A variety of techniques have been developed to scale decision tree classifiers in data mining to extract valuable knowledge. However, these aproaches either cause a loss of accuracy or cannot effectively uncover the data structure. We explore a more promising GA-based decision tree classifier, OOGASC4.5, to integrate the strengths of decision tree algorithms with statistical sampling and genetic algorithm. The proposed program could not only enhance the classification accuracy but assumes the potential advantage of extracting valuable rules as well. The computational results are provided along with analysis and conclusions.
展开▼
