灰色-支持向量回归模型在多应力加速寿命试验寿命预测中的应用

Application of Grey-Support Vector Regression Model in Life Prediction of Multi-stress Accelerated Life Test

针对多应力条件下加速寿命试验中寿命预测模建模困难、小样本情况下预测精度不高、寿命难以估算的问题,建立了灰色-支持向量回归加速模型,在对加速寿命试验数据进行累加生成从而增强规律的基础上进行支持向量回归,从而得到预测值。结合实例对加速模型进行了验证,表明模型具备较好的预测精度和优良的泛化性。

For the problem that the life prediction model is difficult to model in the accelerated life test under multi-stress con⁃ditions,the prediction accuracy is not high in small sample cases,and the life is difficult to estimate,a gray-support vector regres⁃sion prediction model is established,which is generated by gray accumulation of data.Support vector regression operations are per⁃formed on the basis of regularity to obtain predicted values.The model is verified by an example.The results show that the model has high prediction accuracy and good generalization.