自适应模糊决策树算法

Adaptive fuzzy decision tree algorithms

为改善模糊决策树算法凭经验设定参数值的不准确问题,在分析模糊决策树算法的主要参数特征后,提出使用粒子群算法智能设定参数值的自适应模糊决策树算法。实验表明,与经验设定参数值的模糊决策树算法相比,自适应模糊决策树算法生成的模糊决策树的性能明显提高;最后,通过实验数据分析了关键参数之间存在的交互影响关系。

To improve instability caused by the selections of parameters from user＇s demand in fuzzy decision tree algorithms （FDTS）, after discussion the important parameters in FDTS, an adaptive fuzzy decision tree algorithm based on particle swarm optimization is given. As a result, the method can get higher testing accuracy than FDT algorithm with parameters setting from experimental designs. Moreover, the interactive relations between two important parameters also are discussed based on the ex perimental results.