机构地区: 电子科技大学计算机科学与工程学院
出 处: 《电子测量与仪器学报》 2008年第1期10-14,共5页
摘 要: 由于网络故障信息通常表现出不确定、复杂的特点并且对故障特征的描述也多是非完备的,运用支持向量机进行故障诊断经常不能达到其最优性能。本文提出一种新颖的二维bagging集成支持向量机,显著提高了传统支持向量机进行网络故障诊断的精度。实验结果验证了该方法的准确性和有效性。 The collected network fault information always has the feature of uncertainty and complication, and the attributes for describing the fault status are always non-optimal, so that the support vector machines always cannot reach its best performance when it is applied to diagnose the network faults. A novel 2D-bagging ensemble method of support vector machines is proposed in for network faults. Experimental results are provi method. this paper, which can improve the diagnosis accuracy significantly ded to show the excellent accuracy and effectivity of the proposed method.