针对支持向量机(Support vector machine, SVM)的惩罚系数难以确定、核函数必须满足Mercer定理等问题，相关向量机(Relevance vector machine, RVM)应运而生以解决上述问题，并在趋势预测等领域得到一定的应用。核函数是决定RVM预测精度的关键因素之一，目前的研究通常是人为选择单一核函数，因此增加了对参数的依赖性并降低了RVM预测的鲁棒性。为了解决以上问题，提出一种新的自适应多核组合RVM预测方法。该方法首先选择多个核函数，利用粒子滤波产生核函数权重，建立多核组合RVM集，然后经过不断地迭代预测、权值更新和重采样，自适应获取最优多核组合RVM，从而自适应融合多个核函数的特性，克服基于单一核函数RVM的局限，提高预测精度和鲁棒性。利用仿真对提出方法进行了验证，并将其应用于机械设备的剩余寿命预测，取得了比基于单一核函数RVM更好的预测效果。
雷亚国
,
陈吴
,
李乃鹏
,
林京
. 自适应多核组合相关向量机预测方法及其在机械设备剩余寿命预测中的应用[J]. 机械工程学报, 2016
, 52(1)
: 87
-93
.
DOI: 10.3901/JME.2016.01.087
In view of some shortcomings of support vector machine, for instance, it is difficult to select the regularization parameter and the kernel function must satisfy Mercer’s condition, relevance vector machine (RVM) is developed and applied to the field of trend prediction. The performance of RVM, to a large extent, depends on the kernel function. However, a single kernel function is generally selected artificially and subjectively in current studies on RVM, which increases its dependency of the RVM to parameters and decreases the robustness in prediction process. To solve the problem, a new adaptive multi-kernel RVM is proposed for prediction. In the method, multiple kernel functions are selected originally and their weights are generated by the particle filter (PF) algorithm to construct multi-kernel RVM models. Then the optimal multi-kernel RVM model is obtained by iterative processes, i.e., predicting, weights updating and resampling. The effectiveness of the proposed method is validated by a simulation study and a case study of remaining useful life prediction of machinery. The results demonstrate that the proposed method obtains higher prediction accuracies compared with the single kernel RVM models.
[1] 雷亚国，何正嘉，訾艳阳. 基于混合智能新模型的故障诊断[J]. 机械工程学报，2008，44(7)：112-117. LEI Yaguo，HE Zhengjia，ZI Yanyang. Fault diagnosis based on novel hybrid intelligent model[J]. Chinese Journal of Mechanical Engineering，2008，44(7)：112-117.
[2] SOUALHI A，MEDJAHER K，ZERHOUNI N. Bearing health monitoring based on Hilbert–Huang transform，support vector machine, and regression[J]. IEEE Transactions on Instrumentation and Measurement，2015，64(1)：52-62.
[3] NIZAM M，MOHAMED A，HUSSAIN A. Dynamic voltage collapse prediction in power systems using support vector regression[J]. Expert Systems with Applications，2010，37(5)：3730-3736.
[4] TRAN V T，THOM PHAM H，YANG B S，et al. Machine performance degradation assessment and remaining useful life prediction using proportional hazard model and support vector machine[J]. Mechanical Systems and Signal Processing，2012，32：320-330.
[5] 申中杰，陈雪峰，何正嘉，等. 基于相对特征和多变量支持向量机的滚动轴承剩余寿命预测[J]. 机械工程学报，2013，49(2)：183-189. SHEN Zhongjie，CHEN Xuefeng，HE Zhengjia，et al. Remaining life predictions of rolling bearing based on relative features and multivariable support vector machine[J]. Journal of Mechanical Engineering，2013，49(2)：183-189.
[6] TIPPING M E. The relevance vector machine[J]. Advances in Neural Information Processing Systems，2000，12：652-658.
[7] TIPPING M E. Sparse Bayesian learning and the relevance vector machine[J]. The Journal of Machine Learning Research，2001，1：211-244.
[8] CAESARENDRA W，WIDODO A，YANG B S. Application of relevance vector machine and logistic regression for machine degradation assessment[J]. Mechanical Systems and Signal Processing，2010，24(4)：1161-1171.
[9] WIDODO A，YANG B S. Application of relevance vector machine and survival probability to machine degradation assessment[J]. Expert Systems with Applications，2011，38(3)：2592-2599.
[10] 段青，赵建国，马艳. 优化组合核函数相关向量机电力负荷预测模型[J]. 电机与控制学报，2010，14(6)：33-38. DUAN Qing，ZHAO Jianguo，MA Yan. Relevance vector machine based on particle swarm optimization of compounding kernels in electricity load forecasting[J]. Electric Machines and Control，2010，14(6)：33-38.
[11] 朱永利，尹金良. 组合核相关向量机在电力变压器故障诊断中的应用研究[J]. 中国电机工程学报，2013，33(22)：68-74. ZHU Yongli，YIN Jinliang. Study on application of multi-kernel learning relevance vector machines in fault diagnosis of power transformers[J]. Proceedings of the CSEE，2013，33(22)：68-74.
[12] DOUCET A，DE FREITAS N，GORDON N. Sequential Monte Carlo methods in practice[M]. New York：Springer，2001.
[13] SMITS G F，JORDAAN E M. Improved SVM regression using mixtures of kernels[C]//IEEE International Joint Conference on Neural Networks，May 12-17，2002，Honolulu，HI. IEEE，2002，3：2785-2790.
[14] MüLLER K R，SMOLA A J，R?TSCH G，et al. Predicting time series with support vector machines[M]. Berlin Heidelberg：Springer，1997.
[15] ECTOUX P，GOURIVEAU R，MEDJAHER K，et al. PRONOSTIA：An experimental platform for bearings accelerated degradation tests[C]//IEEE International Conference on Prognostics and Health Management，June 18-21，2012，Denver，Colorado，United States. IEEE，2012：1-8.
[16] The IEEE Reliability Society. IEEE PHM 2012 prognostic challenge. outline，experiments，scoring of results，winners[EB/OL]. [2015-03-20]. http：//www.femto-st.fr/f/d/IEEEPHM2012-challenge-details. pdf.
[17] LI Naipeng，LEI Yaguo，LIU Zongyao，et al. A particle filtering-based approach for remaining useful life predication of rolling element bearings[C]//IEEE International Conference on Prognostics and Health Management，June 22-25，2014，Cheney，Washington. IEEE，2014：1-8.
