摘要
在总结了几种比较常见的故障预测方法的基础上，介绍了基于统计学习理论的支持向量回归算法。提出将智能遗传算法用来对支持向量回归模型的参数进行优化选取，并详细介绍了模型参数的选取过程，避免了参数的盲目设置。将建立起来的模型应用于雷达智能BIT故障预测领域，并以一组智能BIT状态监测的数据对预测模型进行训练和验证，实验结果表明支持向量回归模型能有效地对雷达故障进行预测。
On the basis of summarizing several kinds of common failure predication methods, the algorithm of Support Vector Regression （SVR） based on statistical theory is introdueed. To avoid the blind establishment of the parameters, this study proposes intelligent genetie algorithms for optimizing the SVR＇ s parameters and presents the process of seleeting SVR＇ s parameters particularly. Then the SVR model is applied to the radar intelligent Built-in Test failure predieation field. Finally, we adopt the intelligent Built - in Test condition monitoring data to verify the SVR model. The experimental result demonstrated that SVR model ean prediet the radar fault effeetively.
出处
《军械工程学院学报》
2009年第4期49-52,共4页
Journal of Ordnance Engineering College
基金
项目来源：军队科研计划项目
关键词
智能BIT
故障预测
支持向量回归
智能遗传算法
intelligent buih-in test
failure predication
support vector regression
intelligent genetie algorithms
