基于支持向量机的组合分类方法及应用

Ensemble of classification methods based on SVM and its application in diagnosis

为了解决采用神经网络、决策树作为弱分类器的AdaBoost组合分类存在的不足,进一步改善组合分类效果,提出采用支持向量机(SVM)作为弱分类器的一种新的组合分类诊断方法——AdaBoost-SVM。该方法没有采用一个固定的SVM的核参数,而是自适应调整SVM中的核参数,从而得到一组有效的SVM弱分类器。通过对基准数据库的测试及航空发动机故障样本的诊断,结果表明,所提AdaBoost-SVM方法较好地解决了现有的Ada-Boost组合分类方法中存在的弱分类器本身参数选取困难问题及训练轮数的合理选取问题,并具有更好的泛化性能,更适合对分散程度较大、聚类性较差的航空发动机故障样本进行分类。

In order to solve the shortage problem of ensemble of classification using neural networks and decision trees as weak learner and improve the effect of ensemble of classification, a novel approach of classification ensemble named AdaBoost-SVM is presented, which uses SVM as weak learner for AdaBoost. To obtain a set of effective SVM weak learner, this algorithm adaptively adjusts the kernel parameter in SVM instead of using a fixed one. The practical applications in UCI repository and aeorengine faulty samples show that the proposed method solves the problem of selection difficuty for weak learner par rameter and learning cycles in the existing AdaBoost methods and it has better generalization performance and is more fitting to classify the faulty samples scattered greatly,