摘要
针对电子设备的状态预测问题,提出一种基于自适应核学习相关向量机的在线状态预测方法。所提方法将电子设备的状态预测视为一个有监督回归问题。首先通过设备离线数据的后验概率选择最适合预测问题的基函数,然后采用离线数据训练aRVM预测模型,最后在预测过程中通过对在线数据的增样学习来动态更新模型参数。仿真预测实例与实例预测实验的结果表明:所提方法相比传统方法显著提高了预测精度和在线训练效率。
Aimed at the condition prediction of electronic equipment, an online prediction method based on relevance vector machine with adaptive kernel learning（aRVM） is proposed. In the proposed method, the condition prediction of electronic equipment is formulated as a supervised-learning problem. Firstly, the basis function which is most suitable for prediction problem is selected according to the posterior probability of equipment offline data. Then the prediction model based on aRVM is trained by offline data. Finally, the parameters of the model are updated dynamically through incremental training of online samples in the prediction process. Experimental results of both simulation experiment and practical experiment indicate that the proposed method obviously outperforms the traditional one in both prediction accuracy and online training efficiency.
出处
《兵器装备工程学报》
CAS
2017年第11期108-113,共6页
Journal of Ordnance Equipment Engineering
基金
装备预研基金资助项目(9140A27020214JB14435)
关键词
电子设备
状态预测
相关向量机
在线训练
自适应核学习
electronic equipment
status prediction
relevance vector machine
online training
adaptivekernel
