A comparative analysis of methods for pruning decision trees

摘要: In this paper, we address the problem of retrospectively pruning decision trees induced from data, according to a top-down approach. This has received considerable attention in areas pattern recognition and machine learning, many distinct methods have been proposed literature. We make comparative study six well-known with aim understanding their theoretical foundations, computational complexity, strengths weaknesses formulation. Comments on characteristics each method are empirically supported. particular, wide experimentation performed several data sets leads us opposite conclusions predictive accuracy simplified some drawn attribute divergence differences experimental designs. Finally, prove use property reduced error obtain an objective evaluation tendency overprune/underprune observed method.