基于SMO-SVM算法的变压器故障诊断

支持向量机是一种基于统计学理论的机器学习算法,它能在训练样本很少的情况下达到很好的分类效果。针对变压器的特性,提出了以RBF为核函数的非线性支持向量机的二叉树多分类变压器故障诊断模型,利用序贯最优化算法(SMO)对样本进行训练,准确率较高。试验结果表明,SMO-SVM在变压器故障诊断中具有很大的应用潜力。

Fault Diagnosis of Transformer Based on SMO-SVM Algorithm

Support vector machine is a machine based on statistical learning theory learning algorithm,it can in rare cases the training samples to achieve good classification results.The characteristics of transformer is proposed to the nonlinear RBF as kernel function support vector machine multi-classification of the binary tree transformer fault diagnosis model.Use of sequential optimization algorithm(SMO) of the training samples,high accuracy.The results show that,SMO-SVM in fault diagnosis of transformer has gre...