为了提升集成网络的泛化性能,在Boosting或Bagging算法对样本进行扰动的基础上,通过粗糙集约简实现特征属性选择,将样本扰动和输入属性扰动结合起来,提出了Rough_Boosting和Rough_Bagging算法.该算法通过生成精确度高且差异度大的个体网络,提高了集成的泛化能力.实验结果表明,该算法泛化能力明显优于Boosting和Bagging算法,且生成的个体网络差异度更大,与同类算法相比,具有相近或相当的性能. In order to improve the generalization performance of the integrated network, based on the perturbation of the samples by Boosting or Bagging algorithm, the rough set reduction is used to realize the feature attribute selection, and the sample perturbation and the input attribute perturbation are combined to propose Rough_Boosting and Rough_Bagging algorithm. The result shows that the generalization ability of this algorithm is obviously better than that of Boosting and Bagging algorithm, and the individual network generated by the algorithm has more difference, Compared with the algorithm, with similar or equivalent performance.