摘要
在支持向量机理论的基础上,打破了支持向量机的二类辨识传统,引入了基于支持向量机的一类辨识理论,以它为基础设计了基于一类辨识的一类分类器,并将它运用到航空发动机故障诊断中。通过对几种典型故障的分析,证明了该方法的有效性。
Based on SVM theory, a new method named one-class identification was presented in this paper. And the new method broke two-class classification tradition of SVM. Successful application has been achieved to detect several typical fault of aero-engine. The results show that the one class classification can provide a new effective technology to reveal nonstationary fault of aero-engine.
出处
《燃气涡轮试验与研究》
2006年第2期30-32,共3页
Gas Turbine Experiment and Research
